diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..a247b61
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+SOURCES/sos-4.2.tar.gz
+SOURCES/sos-audit-0.3.tgz
diff --git a/.sos.metadata b/.sos.metadata
new file mode 100644
index 0000000..054c91f
--- /dev/null
+++ b/.sos.metadata
@@ -0,0 +1,2 @@
+fe82967b0577076aac104412a9fe35cdb444bde4 SOURCES/sos-4.2.tar.gz
+9d478b9f0085da9178af103078bbf2fd77b0175a SOURCES/sos-audit-0.3.tgz
diff --git a/SOURCES/sos-bz1869561-cpuX-individual-sizelimits.patch b/SOURCES/sos-bz1869561-cpuX-individual-sizelimits.patch
new file mode 100644
index 0000000..4d579d7
--- /dev/null
+++ b/SOURCES/sos-bz1869561-cpuX-individual-sizelimits.patch
@@ -0,0 +1,48 @@
+From b09ed75b09075d86c184b0a63cce9260f2cee4ca Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 30 Aug 2021 11:27:48 +0200
+Subject: [PATCH] [processor] Apply sizelimit to /sys/devices/system/cpu/cpuX
+
+Copy /sys/devices/system/cpu/cpuX with separately applied sizelimit.
+
+This is required for systems with tens/hundreds of CPUs where the
+cumulative directory size exceeds 25MB or even 100MB.
+
+Resolves: #2639
+Closes: #2665
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/processor.py | 9 ++++++++-
+ 1 file changed, 8 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/processor.py b/sos/report/plugins/processor.py
+index 0ddfd126..2df2dc9a 100644
+--- a/sos/report/plugins/processor.py
++++ b/sos/report/plugins/processor.py
+@@ -7,6 +7,7 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ from sos.report.plugins import Plugin, IndependentPlugin
++import os
+
+
+ class Processor(Plugin, IndependentPlugin):
+@@ -34,7 +35,13 @@ class Processor(Plugin, IndependentPlugin):
+ self.add_copy_spec([
+ "/proc/cpuinfo",
+ "/sys/class/cpuid",
+- "/sys/devices/system/cpu"
++ ])
++ # copy /sys/devices/system/cpu/cpuX with separately applied sizelimit
++ # this is required for systems with tens/hundreds of CPUs where the
++ # cumulative directory size exceeds 25MB or even 100MB.
++ cdirs = self.listdir('/sys/devices/system/cpu')
++ self.add_copy_spec([
++ os.path.join('/sys/devices/system/cpu', cdir) for cdir in cdirs
+ ])
+
+ self.add_cmd_output([
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011507-foreman-puma-status.patch b/SOURCES/sos-bz2011507-foreman-puma-status.patch
new file mode 100644
index 0000000..2a80571
--- /dev/null
+++ b/SOURCES/sos-bz2011507-foreman-puma-status.patch
@@ -0,0 +1,69 @@
+From 5a9458d318302c1caef862a868745fc8bdf5c741 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 4 Oct 2021 15:52:36 +0200
+Subject: [PATCH] [foreman] Collect puma status and stats
+
+Collect foreman-puma-status and 'pumactl [gc-|]stats', optionally using
+SCL (if detected).
+
+Resolves: #2712
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/foreman.py | 21 ++++++++++++++++++++-
+ 1 file changed, 20 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/foreman.py b/sos/report/plugins/foreman.py
+index 4539f12b..351794f4 100644
+--- a/sos/report/plugins/foreman.py
++++ b/sos/report/plugins/foreman.py
+@@ -13,6 +13,7 @@ from sos.report.plugins import (Plugin,
+ UbuntuPlugin)
+ from pipes import quote
+ from re import match
++from sos.utilities import is_executable
+
+
+ class Foreman(Plugin):
+@@ -26,7 +27,9 @@ class Foreman(Plugin):
+ option_list = [
+ ('months', 'number of months for dynflow output', 'fast', 1),
+ ('proxyfeatures', 'collect features of smart proxies', 'slow', False),
++ ('puma-gc', 'collect Puma GC stats', 'fast', False),
+ ]
++ pumactl = 'pumactl %s -S /usr/share/foreman/tmp/puma.state'
+
+ def setup(self):
+ # for external DB, search in /etc/foreman/database.yml for:
+@@ -134,6 +138,17 @@ class Foreman(Plugin):
+ suggest_filename='dynflow_sidekiq_status')
+ self.add_journal(units="dynflow-sidekiq@*")
+
++ # Puma stats & status, i.e. foreman-puma-stats, then
++ # pumactl stats -S /usr/share/foreman/tmp/puma.state
++ # and optionally also gc-stats
++ # if on RHEL with Software Collections, wrap the commands accordingly
++ if self.get_option('puma-gc'):
++ self.add_cmd_output(self.pumactl % 'gc-stats',
++ suggest_filename='pumactl_gc-stats')
++ self.add_cmd_output(self.pumactl % 'stats',
++ suggest_filename='pumactl_stats')
++ self.add_cmd_output('/usr/sbin/foreman-puma-status')
++
+ # collect tables sizes, ordered
+ _cmd = self.build_query_cmd(
+ "SELECT table_name, pg_size_pretty(total_bytes) AS total, "
+@@ -297,6 +312,10 @@ class RedHatForeman(Foreman, RedHatPlugin):
+ self.add_file_tags({
+ '/usr/share/foreman/.ssh/ssh_config': 'ssh_foreman_config',
+ })
++ # if we are on RHEL7 with scl, wrap some Puma commands by
++ # scl enable tfm 'command'
++ if self.policy.dist_version() == 7 and is_executable('scl'):
++ self.pumactl = "scl enable tfm '%s'" % self.pumactl
+
+ super(RedHatForeman, self).setup()
+
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011533-unpackaged-recursive-symlink.patch b/SOURCES/sos-bz2011533-unpackaged-recursive-symlink.patch
new file mode 100644
index 0000000..35cc89d
--- /dev/null
+++ b/SOURCES/sos-bz2011533-unpackaged-recursive-symlink.patch
@@ -0,0 +1,42 @@
+From e2ca3d02f36c0db4efaacfb2c1b7d502f38e371c Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 30 Aug 2021 10:18:29 +0200
+Subject: [PATCH] [unpackaged] deal with recursive loop of symlinks properly
+
+When the plugin processes a recursive loop of symlinks, it currently
+hangs in an infinite loop trying to follow the symlinks. Use
+pathlib.Path.resolve() method to return the target directly.
+
+Resolves: #2664
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/unpackaged.py | 5 +++--
+ 1 file changed, 3 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/unpackaged.py b/sos/report/plugins/unpackaged.py
+index e5cc6191..9d68077c 100644
+--- a/sos/report/plugins/unpackaged.py
++++ b/sos/report/plugins/unpackaged.py
+@@ -10,6 +10,7 @@ from sos.report.plugins import Plugin, RedHatPlugin
+
+ import os
+ import stat
++from pathlib import Path
+
+
+ class Unpackaged(Plugin, RedHatPlugin):
+@@ -41,8 +42,8 @@ class Unpackaged(Plugin, RedHatPlugin):
+ for name in files:
+ path = os.path.join(root, name)
+ try:
+- while stat.S_ISLNK(os.lstat(path).st_mode):
+- path = os.path.abspath(os.readlink(path))
++ if stat.S_ISLNK(os.lstat(path).st_mode):
++ path = Path(path).resolve()
+ except Exception:
+ continue
+ file_list.append(os.path.realpath(path))
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011534-opacapture-under-allow-system-changes.patch b/SOURCES/sos-bz2011534-opacapture-under-allow-system-changes.patch
new file mode 100644
index 0000000..39f9c8a
--- /dev/null
+++ b/SOURCES/sos-bz2011534-opacapture-under-allow-system-changes.patch
@@ -0,0 +1,49 @@
+From 66ebb8256b1326573cbcb2d134545635dfead3bc Mon Sep 17 00:00:00 2001
+From: Jose Castillo <jcastillo@redhat.com>
+Date: Sun, 29 Aug 2021 15:35:09 +0200
+Subject: [PATCH] [omnipath_client] Ensure opacapture runs only with
+ allow-system-changes
+
+While omnipath_client plugin is collecting "opacapture",
+`depmod -a` command is executed to regenerates some files
+under /usr/lib/modules/$kernel.
+
+modules.dep
+modules.dep.bin
+modules.devname
+modules.softdep
+modules.symbols
+modules.symbols.bin
+
+This patch ensures that the command is only run when
+the option --allow-system-changes is used.
+
+Fixes: RHBZ#1998433
+
+Signed-off-by: Jose Castillo <jcastillo@redhat.com>
+---
+ sos/report/plugins/omnipath_client.py | 9 +++++++--
+ 1 file changed, 7 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/omnipath_client.py b/sos/report/plugins/omnipath_client.py
+index 1ec01384..4e988c5c 100644
+--- a/sos/report/plugins/omnipath_client.py
++++ b/sos/report/plugins/omnipath_client.py
+@@ -45,7 +45,12 @@ class OmnipathClient(Plugin, RedHatPlugin):
+ # rather than storing it somewhere under /var/tmp and copying it via
+ # add_copy_spec, add it directly to sos_commands/<plugin> dir by
+ # building a path argument using self.get_cmd_output_path().
+- self.add_cmd_output("opacapture %s" % join(self.get_cmd_output_path(),
+- "opacapture.tgz"))
++ # This command calls 'depmod -a', so lets make sure we
++ # specified the 'allow-system-changes' option before running it.
++ if self.get_option('allow_system_changes'):
++ self.add_cmd_output("opacapture %s" %
++ join(self.get_cmd_output_path(),
++ "opacapture.tgz"),
++ changes=True)
+
+ # vim: set et ts=4 sw=4 :
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011535-kernel-psi.patch b/SOURCES/sos-bz2011535-kernel-psi.patch
new file mode 100644
index 0000000..1a9d5e0
--- /dev/null
+++ b/SOURCES/sos-bz2011535-kernel-psi.patch
@@ -0,0 +1,33 @@
+From 23e523b6b9784390c7ce2c5af654ab497fb10aaf Mon Sep 17 00:00:00 2001
+From: Jose Castillo <jcastillo@redhat.com>
+Date: Wed, 8 Sep 2021 09:25:24 +0200
+Subject: [PATCH] [kernel] Capture Pressure Stall Information
+
+Kernel 4.20 includes PSI metrics for CPU, memeory and IO.
+The feature is enabled after adding "psi=1" as
+kernel boot parameter.
+The information is captured in files
+in the directory /proc/pressure.
+
+Signed-off-by: Jose Castillo <jcastillo@redhat.com>
+---
+ sos/report/plugins/kernel.py | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/kernel.py b/sos/report/plugins/kernel.py
+index 8c5e5e11..803f5e30 100644
+--- a/sos/report/plugins/kernel.py
++++ b/sos/report/plugins/kernel.py
+@@ -112,7 +112,8 @@ class Kernel(Plugin, IndependentPlugin):
+ "/sys/kernel/debug/extfrag/unusable_index",
+ "/sys/kernel/debug/extfrag/extfrag_index",
+ clocksource_path + "available_clocksource",
+- clocksource_path + "current_clocksource"
++ clocksource_path + "current_clocksource",
++ "/proc/pressure/"
+ ])
+
+ if self.get_option("with-timer"):
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011536-iptables-based-on-ntf.patch b/SOURCES/sos-bz2011536-iptables-based-on-ntf.patch
new file mode 100644
index 0000000..5ccc61f
--- /dev/null
+++ b/SOURCES/sos-bz2011536-iptables-based-on-ntf.patch
@@ -0,0 +1,303 @@
+From 2ab8ba3ecbd52e452cc554d515e0782801dcb4b6 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 8 Sep 2021 15:31:48 +0200
+Subject: [PATCH] [firewalld] collect nft rules in firewall_tables only
+
+We collect 'nft list ruleset' in both plugins, while:
+- nft is not shipped by firewalld package, so we should not collect
+it in firewalld plugin
+- running the command requires both nf_tables and nfnetlink kmods, so
+we should use both kmods in the predicate
+
+Resolves: #2679
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/firewall_tables.py | 9 +++++----
+ sos/report/plugins/firewalld.py | 8 +-------
+ 2 files changed, 6 insertions(+), 11 deletions(-)
+
+diff --git a/sos/report/plugins/firewall_tables.py b/sos/report/plugins/firewall_tables.py
+index 56058d3bf9..63a7dddeb5 100644
+--- a/sos/report/plugins/firewall_tables.py
++++ b/sos/report/plugins/firewall_tables.py
+@@ -40,10 +40,11 @@ def collect_nftables(self):
+ """ Collects nftables rulesets with 'nft' commands if the modules
+ are present """
+
+- self.add_cmd_output(
+- "nft list ruleset",
+- pred=SoSPredicate(self, kmods=['nf_tables'])
+- )
++ # collect nftables ruleset
++ nft_pred = SoSPredicate(self,
++ kmods=['nf_tables', 'nfnetlink'],
++ required={'kmods': 'all'})
++ self.add_cmd_output("nft list ruleset", pred=nft_pred, changes=True)
+
+ def setup(self):
+ # collect iptables -t for any existing table, if we can't read the
+diff --git a/sos/report/plugins/firewalld.py b/sos/report/plugins/firewalld.py
+index ec83527ed7..9401bfd239 100644
+--- a/sos/report/plugins/firewalld.py
++++ b/sos/report/plugins/firewalld.py
+@@ -9,7 +9,7 @@
+ #
+ # See the LICENSE file in the source distribution for further information.
+
+-from sos.report.plugins import Plugin, RedHatPlugin, SoSPredicate
++from sos.report.plugins import Plugin, RedHatPlugin
+
+
+ class FirewallD(Plugin, RedHatPlugin):
+@@ -35,12 +35,6 @@ def setup(self):
+ "/var/log/firewalld",
+ ])
+
+- # collect nftables ruleset
+- nft_pred = SoSPredicate(self,
+- kmods=['nf_tables', 'nfnetlink'],
+- required={'kmods': 'all'})
+- self.add_cmd_output("nft list ruleset", pred=nft_pred, changes=True)
+-
+ # use a 10s timeout to workaround dbus problems in
+ # docker containers.
+ self.add_cmd_output([
+--
+2.31.1
+
+
+From 2a7cf53b61943907dc823cf893530b620a87946c Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 15 Oct 2021 22:31:36 +0200
+Subject: [PATCH 1/3] [report] Use log_skipped_cmd method inside
+ collect_cmd_output
+
+Also, remove obsolete parameters of the log_skipped_cmd method.
+
+Related: #2724
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/__init__.py | 26 ++++++++------------------
+ 1 file changed, 8 insertions(+), 18 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index ec138f83..b60ab5f6 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -876,8 +876,7 @@ class Plugin():
+ return bool(pred)
+ return False
+
+- def log_skipped_cmd(self, pred, cmd, kmods=False, services=False,
+- changes=False):
++ def log_skipped_cmd(self, cmd, pred, changes=False):
+ """Log that a command was skipped due to predicate evaluation.
+
+ Emit a warning message indicating that a command was skipped due
+@@ -887,21 +886,17 @@ class Plugin():
+ message indicating that the missing data can be collected by using
+ the "--allow-system-changes" command line option will be included.
+
+- :param pred: The predicate that caused the command to be skipped
+- :type pred: ``SoSPredicate``
+-
+ :param cmd: The command that was skipped
+ :type cmd: ``str``
+
+- :param kmods: Did kernel modules cause the command to be skipped
+- :type kmods: ``bool``
+-
+- :param services: Did services cause the command to be skipped
+- :type services: ``bool``
++ :param pred: The predicate that caused the command to be skipped
++ :type pred: ``SoSPredicate``
+
+ :param changes: Is the `--allow-system-changes` enabled
+ :type changes: ``bool``
+ """
++ if pred is None:
++ pred = SoSPredicate(self)
+ msg = "skipped command '%s': %s" % (cmd, pred.report_failure())
+
+ if changes:
+@@ -1700,9 +1693,7 @@ class Plugin():
+ self.collect_cmds.append(soscmd)
+ self._log_info("added cmd output '%s'" % soscmd.cmd)
+ else:
+- self.log_skipped_cmd(pred, soscmd.cmd, kmods=bool(pred.kmods),
+- services=bool(pred.services),
+- changes=soscmd.changes)
++ self.log_skipped_cmd(soscmd.cmd, pred, changes=soscmd.changes)
+
+ def add_cmd_output(self, cmds, suggest_filename=None,
+ root_symlink=None, timeout=None, stderr=True,
+@@ -2112,7 +2103,7 @@ class Plugin():
+ root_symlink=False, timeout=None,
+ stderr=True, chroot=True, runat=None, env=None,
+ binary=False, sizelimit=None, pred=None,
+- subdir=None, tags=[]):
++ changes=False, subdir=None, tags=[]):
+ """Execute a command and save the output to a file for inclusion in the
+ report, then return the results for further use by the plugin
+
+@@ -2163,8 +2154,7 @@ class Plugin():
+ :rtype: ``dict``
+ """
+ if not self.test_predicate(cmd=True, pred=pred):
+- self._log_info("skipped cmd output '%s' due to predicate (%s)" %
+- (cmd, self.get_predicate(cmd=True, pred=pred)))
++ self.log_skipped_cmd(cmd, pred, changes=changes)
+ return {
+ 'status': None, # don't match on if result['status'] checks
+ 'output': '',
+--
+2.31.1
+
+
+From 6b1bea0ffb1df7f8e5001b06cf25f0741b007ddd Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 15 Oct 2021 22:34:01 +0200
+Subject: [PATCH 2/3] [firewall_tables] call iptables -t <table> based on nft
+ list
+
+If iptables are not realy in use, calling iptables -t <table>
+would load corresponding nft table.
+
+Therefore, call iptables -t only for the tables from "nft list ruleset"
+output.
+
+Example: nft list ruleset contains
+
+table ip mangle {
+..
+}
+
+so we can collect iptable -t mangle -nvL .
+
+The same applies to ip6tables as well.
+
+Resolves: #2724
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/firewall_tables.py | 29 ++++++++++++++++++++-------
+ 1 file changed, 22 insertions(+), 7 deletions(-)
+
+diff --git a/sos/report/plugins/firewall_tables.py b/sos/report/plugins/firewall_tables.py
+index 63a7ddde..ef04d939 100644
+--- a/sos/report/plugins/firewall_tables.py
++++ b/sos/report/plugins/firewall_tables.py
+@@ -44,26 +44,41 @@ class firewall_tables(Plugin, IndependentPlugin):
+ nft_pred = SoSPredicate(self,
+ kmods=['nf_tables', 'nfnetlink'],
+ required={'kmods': 'all'})
+- self.add_cmd_output("nft list ruleset", pred=nft_pred, changes=True)
++ return self.collect_cmd_output("nft list ruleset", pred=nft_pred,
++ changes=True)
+
+ def setup(self):
++ # first, collect "nft list ruleset" as collecting commands like
++ # ip6tables -t mangle -nvL
++ # depends on its output
++ # store in nft_ip_tables lists of ip[|6] tables from nft list
++ nft_list = self.collect_nftables()
++ nft_ip_tables = {'ip': [], 'ip6': []}
++ nft_lines = nft_list['output'] if nft_list['status'] == 0 else ''
++ for line in nft_lines.splitlines():
++ words = line.split()[0:3]
++ if len(words) == 3 and words[0] == 'table' and \
++ words[1] in nft_ip_tables.keys():
++ nft_ip_tables[words[1]].append(words[2])
+ # collect iptables -t for any existing table, if we can't read the
+ # tables, collect 2 default ones (mangle, filter)
++ # do collect them only when relevant nft list ruleset exists
++ default_ip_tables = "mangle\nfilter\n"
+ try:
+ ip_tables_names = open("/proc/net/ip_tables_names").read()
+ except IOError:
+- ip_tables_names = "mangle\nfilter\n"
++ ip_tables_names = default_ip_tables
+ for table in ip_tables_names.splitlines():
+- self.collect_iptable(table)
++ if nft_list['status'] == 0 and table in nft_ip_tables['ip']:
++ self.collect_iptable(table)
+ # collect the same for ip6tables
+ try:
+ ip_tables_names = open("/proc/net/ip6_tables_names").read()
+ except IOError:
+- ip_tables_names = "mangle\nfilter\n"
++ ip_tables_names = default_ip_tables
+ for table in ip_tables_names.splitlines():
+- self.collect_ip6table(table)
+-
+- self.collect_nftables()
++ if nft_list['status'] == 0 and table in nft_ip_tables['ip6']:
++ self.collect_ip6table(table)
+
+ # When iptables is called it will load the modules
+ # iptables_filter (for kernel <= 3) or
+--
+2.31.1
+
+
+From 464bd2d2e83f203e369f2ba7671bbb7da53e06f6 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Sun, 24 Oct 2021 16:00:31 +0200
+Subject: [PATCH 3/3] [firewall_tables] Call iptables only when nft ip filter
+ table exists
+
+iptables -vnxL creates nft 'ip filter' table if it does not exist, hence
+we must guard iptables execution by presence of the nft table.
+
+An equivalent logic applies to ip6tables.
+
+Resolves: #2724
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/firewall_tables.py | 26 ++++++++++++++------------
+ 1 file changed, 14 insertions(+), 12 deletions(-)
+
+diff --git a/sos/report/plugins/firewall_tables.py b/sos/report/plugins/firewall_tables.py
+index ef04d939..7eafd60f 100644
+--- a/sos/report/plugins/firewall_tables.py
++++ b/sos/report/plugins/firewall_tables.py
+@@ -80,19 +80,21 @@ class firewall_tables(Plugin, IndependentPlugin):
+ if nft_list['status'] == 0 and table in nft_ip_tables['ip6']:
+ self.collect_ip6table(table)
+
+- # When iptables is called it will load the modules
+- # iptables_filter (for kernel <= 3) or
+- # nf_tables (for kernel >= 4) if they are not loaded.
++ # When iptables is called it will load:
++ # 1) the modules iptables_filter (for kernel <= 3) or
++ # nf_tables (for kernel >= 4) if they are not loaded.
++ # 2) nft 'ip filter' table will be created
+ # The same goes for ipv6.
+- self.add_cmd_output(
+- "iptables -vnxL",
+- pred=SoSPredicate(self, kmods=['iptable_filter', 'nf_tables'])
+- )
+-
+- self.add_cmd_output(
+- "ip6tables -vnxL",
+- pred=SoSPredicate(self, kmods=['ip6table_filter', 'nf_tables'])
+- )
++ if nft_list['status'] != 0 or 'filter' in nft_ip_tables['ip']:
++ self.add_cmd_output(
++ "iptables -vnxL",
++ pred=SoSPredicate(self, kmods=['iptable_filter', 'nf_tables'])
++ )
++ if nft_list['status'] != 0 or 'filter' in nft_ip_tables['ip6']:
++ self.add_cmd_output(
++ "ip6tables -vnxL",
++ pred=SoSPredicate(self, kmods=['ip6table_filter', 'nf_tables'])
++ )
+
+ self.add_copy_spec([
+ "/etc/nftables",
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2011537-estimate-only-option.patch b/SOURCES/sos-bz2011537-estimate-only-option.patch
new file mode 100644
index 0000000..a1a96c4
--- /dev/null
+++ b/SOURCES/sos-bz2011537-estimate-only-option.patch
@@ -0,0 +1,1316 @@
+From 5b245b1e449c6a05d09034bcb8290bffded79327 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 8 Sep 2021 17:04:58 +0200
+Subject: [PATCH] [report] Implement --estimate-only
+
+Add report option --estimate-only to estimate disk space requirements
+when running a sos report.
+
+Resolves: #2673
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ man/en/sos-report.1 | 13 +++++++-
+ sos/report/__init__.py | 74 ++++++++++++++++++++++++++++++++++++++++--
+ 2 files changed, 84 insertions(+), 3 deletions(-)
+
+diff --git a/man/en/sos-report.1 b/man/en/sos-report.1
+index 36b337df..e8efc8f8 100644
+--- a/man/en/sos-report.1
++++ b/man/en/sos-report.1
+@@ -14,7 +14,7 @@ sos report \- Collect and package diagnostic and support data
+ [--preset preset] [--add-preset add_preset]\fR
+ [--del-preset del_preset] [--desc description]\fR
+ [--batch] [--build] [--debug] [--dry-run]\fR
+- [--label label] [--case-id id]\fR
++ [--estimate-only] [--label label] [--case-id id]\fR
+ [--threads threads]\fR
+ [--plugin-timeout TIMEOUT]\fR
+ [--cmd-timeout TIMEOUT]\fR
+@@ -317,6 +317,17 @@ output, or string data from the system. The resulting logs may be used
+ to understand the actions that sos would have taken without the dry run
+ option.
+ .TP
++.B \--estimate-only
++Estimate disk space requirements when running sos report. This can be valuable
++to prevent sosreport working dir to consume all free disk space. No plugin data
++is available at the end.
++
++Plugins will be collected sequentially, size of collected files and commands outputs
++will be calculated and the plugin files will be immediatelly deleted prior execution
++of the next plugin. This still can consume whole free disk space, though. Please note,
++size estimations may not be accurate for highly utilized systems due to changes between
++an estimate and a real execution.
++.TP
+ .B \--upload
+ If specified, attempt to upload the resulting archive to a vendor defined location.
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index 82484f1d..b033f621 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -86,6 +86,7 @@ class SoSReport(SoSComponent):
+ 'desc': '',
+ 'domains': [],
+ 'dry_run': False,
++ 'estimate_only': False,
+ 'experimental': False,
+ 'enable_plugins': [],
+ 'keywords': [],
+@@ -137,6 +138,7 @@ class SoSReport(SoSComponent):
+ self._args = args
+ self.sysroot = "/"
+ self.preset = None
++ self.estimated_plugsizes = {}
+
+ self.print_header()
+ self._set_debug()
+@@ -223,6 +225,11 @@ class SoSReport(SoSComponent):
+ help="Description for a new preset",)
+ report_grp.add_argument("--dry-run", action="store_true",
+ help="Run plugins but do not collect data")
++ report_grp.add_argument("--estimate-only", action="store_true",
++ help="Approximate disk space requirements for "
++ "a real sos run; disables --clean and "
++ "--collect, sets --threads=1 and "
++ "--no-postproc")
+ report_grp.add_argument("--experimental", action="store_true",
+ dest="experimental", default=False,
+ help="enable experimental plugins")
+@@ -700,6 +700,33 @@ class SoSReport(SoSComponent):
+ self.all_options.append((plugin, plugin_name, optname,
+ optparm))
+
++ def _set_estimate_only(self):
++ # set estimate-only mode by enforcing some options settings
++ # and return a corresponding log messages string
++ msg = "\nEstimate-only mode enabled"
++ ext_msg = []
++ if self.opts.threads > 1:
++ ext_msg += ["--threads=%s overriden to 1" % self.opts.threads, ]
++ self.opts.threads = 1
++ if not self.opts.build:
++ ext_msg += ["--build enabled", ]
++ self.opts.build = True
++ if not self.opts.no_postproc:
++ ext_msg += ["--no-postproc enabled", ]
++ self.opts.no_postproc = True
++ if self.opts.clean:
++ ext_msg += ["--clean disabled", ]
++ self.opts.clean = False
++ if self.opts.upload:
++ ext_msg += ["--upload* options disabled", ]
++ self.opts.upload = False
++ if ext_msg:
++ msg += ", which overrides some options:\n " + "\n ".join(ext_msg)
++ else:
++ msg += "."
++ msg += "\n\n"
++ return msg
++
+ def _report_profiles_and_plugins(self):
+ self.ui_log.info("")
+ if len(self.loaded_plugins):
+@@ -875,10 +909,12 @@ class SoSReport(SoSComponent):
+ return True
+
+ def batch(self):
++ msg = self.policy.get_msg()
++ if self.opts.estimate_only:
++ msg += self._set_estimate_only()
+ if self.opts.batch:
+- self.ui_log.info(self.policy.get_msg())
++ self.ui_log.info(msg)
+ else:
+- msg = self.policy.get_msg()
+ msg += _("Press ENTER to continue, or CTRL-C to quit.\n")
+ try:
+ input(msg)
+@@ -1011,6 +1047,22 @@ class SoSReport(SoSComponent):
+ self.running_plugs.remove(plugin[1])
+ self.loaded_plugins[plugin[0]-1][1].set_timeout_hit()
+ pool._threads.clear()
++ if self.opts.estimate_only:
++ from pathlib import Path
++ tmpdir_path = Path(self.archive.get_tmp_dir())
++ self.estimated_plugsizes[plugin[1]] = sum(
++ [f.stat().st_size for f in tmpdir_path.glob('**/*')
++ if (os.path.isfile(f) and not os.path.islink(f))])
++ # remove whole tmp_dir content - including "sos_commands" and
++ # similar dirs that will be re-created on demand by next plugin
++ # if needed; it is less error-prone approach than skipping
++ # deletion of some dirs but deleting their content
++ for f in os.listdir(self.archive.get_tmp_dir()):
++ f = os.path.join(self.archive.get_tmp_dir(), f)
++ if os.path.isdir(f):
++ rmtree(f)
++ else:
++ os.unlink(f)
+ return True
+
+ def collect_plugin(self, plugin):
+@@ -1330,6 +1382,24 @@ class SoSReport(SoSComponent):
+ self.policy.display_results(archive, directory, checksum,
+ map_file=map_file)
+
++ if self.opts.estimate_only:
++ from sos.utilities import get_human_readable
++ _sum = get_human_readable(sum(self.estimated_plugsizes.values()))
++ self.ui_log.info("Estimated disk space requirement for whole "
++ "uncompressed sos report directory: %s" % _sum)
++ bigplugins = sorted(self.estimated_plugsizes.items(),
++ key=lambda x: x[1], reverse=True)[:3]
++ bp_out = ", ".join("%s: %s" %
++ (p, get_human_readable(v, precision=0))
++ for p, v in bigplugins)
++ self.ui_log.info("Three biggest plugins: %s" % bp_out)
++ self.ui_log.info("")
++ self.ui_log.info("Please note the estimation is relevant to the "
++ "current options.")
++ self.ui_log.info("Be aware that the real disk space requirements "
++ "might be different.")
++ self.ui_log.info("")
++
+ if self.opts.upload or self.opts.upload_url:
+ if not self.opts.build:
+ try:
+--
+2.31.1
+
+From 7ae47e6c0717c0b56c3368008dd99a87f7f436d5 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 13 Oct 2021 20:21:16 +0200
+Subject: [PATCH] [report] Count with sos_logs and sos_reports in
+ --estimate-only
+
+Currently, we estimate just plugins' disk space and ignore sos_logs
+or sos_reports directories - although they can occupy nontrivial disk
+space as well.
+
+Resolves: #2723
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/__init__.py | 8 ++++++++
+ 1 file changed, 8 insertions(+)
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index e35c7e8d..7feb31ee 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -1380,6 +1380,14 @@ class SoSReport(SoSComponent):
+
+ if self.opts.estimate_only:
+ from sos.utilities import get_human_readable
++ from pathlib import Path
++ # add sos_logs, sos_reports dirs, etc., basically everything
++ # that remained in self.tmpdir after plugins' contents removal
++ # that still will be moved to the sos report final directory path
++ tmpdir_path = Path(self.tmpdir)
++ self.estimated_plugsizes['sos_logs_reports'] = sum(
++ [f.stat().st_size for f in tmpdir_path.glob('**/*')])
++
+ _sum = get_human_readable(sum(self.estimated_plugsizes.values()))
+ self.ui_log.info("Estimated disk space requirement for whole "
+ "uncompressed sos report directory: %s" % _sum)
+--
+2.31.1
+
+From 4293f3317505661e8f32ba94ad87310996fa1626 Mon Sep 17 00:00:00 2001
+From: Eric Desrochers <eric.desrochers@canonical.com>
+Date: Tue, 19 Oct 2021 12:18:40 -0400
+Subject: [PATCH] [report] check for symlink before rmtree when opt
+ estimate-only is use
+
+Check if the dir is also symlink before performing rmtree()
+method so that unlink() method can be used instead.
+
+Traceback (most recent call last):
+ File "./bin/sos", line 22, in <module>
+ sos.execute()
+ File "/tmp/sos/sos/__init__.py", line 186, in execute
+ self._component.execute()
+OSError: Cannot call rmtree on a symbolic link
+
+Closes: #2727
+
+Signed-off-by: Eric Desrochers <eric.desrochers@canonical.com>
+---
+ sos/report/__init__.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index 7feb31ee..1b5bc97d 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -1059,7 +1059,7 @@ class SoSReport(SoSComponent):
+ # deletion of some dirs but deleting their content
+ for f in os.listdir(self.archive.get_tmp_dir()):
+ f = os.path.join(self.archive.get_tmp_dir(), f)
+- if os.path.isdir(f):
++ if os.path.isdir(f) and not os.path.islink(f):
+ rmtree(f)
+ else:
+ os.unlink(f)
+--
+2.31.1
+
+From 589d47c93257b55bc796ef6ac25b88c974ee3d72 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 8 Nov 2021 16:38:24 +0100
+Subject: [PATCH] [report] Calculate sizes of dirs, symlinks and manifest in
+ estimate mode
+
+Enhance --estimate-mode to calculate sizes of also:
+- symlinks
+- directories themselves
+- manifest.json file
+
+Use os.lstat() method instead of os.stat() to properly calculate the
+sizes (and not destinations of symlinks, e.g.).
+
+Print five biggest plugins instead of three as sos logs and reports do
+stand as one "plugin" in the list, often.
+
+Resolves: #2752
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/__init__.py | 56 +++++++++++++++++++++---------------------
+ 1 file changed, 28 insertions(+), 28 deletions(-)
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index 10952566..a4c92acc 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -1050,8 +1050,7 @@ class SoSReport(SoSComponent):
+ from pathlib import Path
+ tmpdir_path = Path(self.archive.get_tmp_dir())
+ self.estimated_plugsizes[plugin[1]] = sum(
+- [f.stat().st_size for f in tmpdir_path.glob('**/*')
+- if (os.path.isfile(f) and not os.path.islink(f))])
++ [f.lstat().st_size for f in tmpdir_path.glob('**/*')])
+ # remove whole tmp_dir content - including "sos_commands" and
+ # similar dirs that will be re-created on demand by next plugin
+ # if needed; it is less error-prone approach than skipping
+@@ -1273,6 +1272,33 @@ class SoSReport(SoSComponent):
+ short_name='manifest.json'
+ )
+
++ # print results in estimate mode (to include also just added manifest)
++ if self.opts.estimate_only:
++ from sos.utilities import get_human_readable
++ from pathlib import Path
++ # add sos_logs, sos_reports dirs, etc., basically everything
++ # that remained in self.tmpdir after plugins' contents removal
++ # that still will be moved to the sos report final directory path
++ tmpdir_path = Path(self.tmpdir)
++ self.estimated_plugsizes['sos_logs_reports'] = sum(
++ [f.lstat().st_size for f in tmpdir_path.glob('**/*')])
++
++ _sum = get_human_readable(sum(self.estimated_plugsizes.values()))
++ self.ui_log.info("Estimated disk space requirement for whole "
++ "uncompressed sos report directory: %s" % _sum)
++ bigplugins = sorted(self.estimated_plugsizes.items(),
++ key=lambda x: x[1], reverse=True)[:5]
++ bp_out = ", ".join("%s: %s" %
++ (p, get_human_readable(v, precision=0))
++ for p, v in bigplugins)
++ self.ui_log.info("Five biggest plugins: %s" % bp_out)
++ self.ui_log.info("")
++ self.ui_log.info("Please note the estimation is relevant to the "
++ "current options.")
++ self.ui_log.info("Be aware that the real disk space requirements "
++ "might be different.")
++ self.ui_log.info("")
++
+ # package up and compress the results
+ if not self.opts.build:
+ old_umask = os.umask(0o077)
+@@ -1377,32 +1403,6 @@ class SoSReport(SoSComponent):
+ self.policy.display_results(archive, directory, checksum,
+ map_file=map_file)
+
+- if self.opts.estimate_only:
+- from sos.utilities import get_human_readable
+- from pathlib import Path
+- # add sos_logs, sos_reports dirs, etc., basically everything
+- # that remained in self.tmpdir after plugins' contents removal
+- # that still will be moved to the sos report final directory path
+- tmpdir_path = Path(self.tmpdir)
+- self.estimated_plugsizes['sos_logs_reports'] = sum(
+- [f.stat().st_size for f in tmpdir_path.glob('**/*')])
+-
+- _sum = get_human_readable(sum(self.estimated_plugsizes.values()))
+- self.ui_log.info("Estimated disk space requirement for whole "
+- "uncompressed sos report directory: %s" % _sum)
+- bigplugins = sorted(self.estimated_plugsizes.items(),
+- key=lambda x: x[1], reverse=True)[:3]
+- bp_out = ", ".join("%s: %s" %
+- (p, get_human_readable(v, precision=0))
+- for p, v in bigplugins)
+- self.ui_log.info("Three biggest plugins: %s" % bp_out)
+- self.ui_log.info("")
+- self.ui_log.info("Please note the estimation is relevant to the "
+- "current options.")
+- self.ui_log.info("Be aware that the real disk space requirements "
+- "might be different.")
+- self.ui_log.info("")
+-
+ if self.opts.upload or self.opts.upload_url:
+ if not self.opts.build:
+ try:
+--
+2.31.1
+
+From c6a5bbb8d75aadd5c7f76d3f469929aba2cf8060 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 5 Jan 2022 10:33:58 +0100
+Subject: [PATCH] [report] Provide better warning about estimate-mode
+
+As --estimate-only calculates disk usage based on `stat` data that
+differs from outputs of other commands like `du`, enhance the warning
+about reliability of the calculated estimation.
+
+Also add a rule-of-thumb recommendation of real disk space requirements.
+
+Resolves: #2815
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ man/en/sos-report.1 | 10 +++++++---
+ sos/report/__init__.py | 3 ++-
+ 2 files changed, 9 insertions(+), 4 deletions(-)
+
+diff --git a/man/en/sos-report.1 b/man/en/sos-report.1
+index 464a77e54..e34773986 100644
+--- a/man/en/sos-report.1
++++ b/man/en/sos-report.1
+@@ -343,9 +343,13 @@ is available at the end.
+
+ Plugins will be collected sequentially, size of collected files and commands outputs
+ will be calculated and the plugin files will be immediatelly deleted prior execution
+-of the next plugin. This still can consume whole free disk space, though. Please note,
+-size estimations may not be accurate for highly utilized systems due to changes between
+-an estimate and a real execution.
++of the next plugin. This still can consume whole free disk space, though.
++
++Please note, size estimations may not be accurate for highly utilized systems due to
++changes between an estimate and a real execution. Also some difference between
++estimation (using `stat` command) and other commands used (i.e. `du`).
++
++A rule of thumb is to reserve at least double the estimation.
+ .TP
+ .B \--upload
+ If specified, attempt to upload the resulting archive to a vendor defined location.
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index ef61fb344..e0617b45e 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -1330,7 +1330,8 @@ def final_work(self):
+ self.ui_log.info("Please note the estimation is relevant to the "
+ "current options.")
+ self.ui_log.info("Be aware that the real disk space requirements "
+- "might be different.")
++ "might be different. A rule of thumb is to "
++ "reserve at least double the estimation.")
+ self.ui_log.info("")
+
+ # package up and compress the results
+From f22efe044f1f0565b57d6aeca2081a5227e0312c Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 14 Feb 2022 09:37:30 -0500
+Subject: [PATCH] [utilities] Don't try to chroot to /
+
+With the recent fix for sysroot being `None` to always being (correctly)
+`/`, we should guard against situations where `sos_get_command_output()`
+would now try to chroot to `/` before running any command. Incidentally,
+this would also cause our unittests to fail if they were run by a
+non-root user.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/utilities.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/utilities.py b/sos/utilities.py
+index 6b13415b..d782123a 100644
+--- a/sos/utilities.py
++++ b/sos/utilities.py
+@@ -120,7 +120,7 @@ def sos_get_command_output(command, timeout=TIMEOUT_DEFAULT, stderr=False,
+ # closure are caught in the parent (chroot and chdir are bound from
+ # the enclosing scope).
+ def _child_prep_fn():
+- if (chroot):
++ if chroot and chroot != '/':
+ os.chroot(chroot)
+ if (chdir):
+ os.chdir(chdir)
+--
+2.34.1
+From 3d064102f8ca6662fd9602512e1cb05cf8746dfd Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 27 Sep 2021 19:01:16 -0400
+Subject: [PATCH] [Systemd, Policy] Correct InitSystem chrooting when chroot is
+ needed
+
+This commit resolves a situation in which `sos` is being run in a
+container but the `SystemdInit` InitSystem would not properly load
+information from the host, thus causing the `Plugin.is_service*()`
+methods to erroneously fail or return `False`.
+
+Fix this scenario by pulling the `_container_init()` and related logic
+to check for a containerized host sysroot out of the Red Hat specific
+policy and into the base `LinuxPolicy` class so that the init system can
+be initialized with the correct sysroot, which is now used to chroot the
+calls to the relevant `systemctl` commands.
+
+For now, this does impose the use of looking for the `container` env var
+(automatically set by docker, podman, and crio regardless of
+distribution) and the use of the `HOST` env var to read where the host's
+`/` filesystem is mounted within the container. If desired in the
+future, this can be changed to allow policy-specific overrides. For now
+however, this extends host collection via an sos container for all
+distributions currently shipping sos.
+
+Note that this issue only affected the `InitSystem` abstraction for
+loading information about local services, and did not affect init system
+related commands called by plugins as part of those collections.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/policies/distros/__init__.py | 28 ++++++++++++++++++++++++++-
+ sos/policies/distros/redhat.py | 27 +-------------------------
+ sos/policies/init_systems/__init__.py | 13 +++++++++++--
+ sos/policies/init_systems/systemd.py | 7 ++++---
+ 4 files changed, 43 insertions(+), 32 deletions(-)
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index f5b9fd5b01..c33a356a75 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -29,6 +29,10 @@
+ except ImportError:
+ REQUESTS_LOADED = False
+
++# Container environment variables for detecting if we're in a container
++ENV_CONTAINER = 'container'
++ENV_HOST_SYSROOT = 'HOST'
++
+
+ class LinuxPolicy(Policy):
+ """This policy is meant to be an abc class that provides common
+@@ -69,10 +73,17 @@ def __init__(self, sysroot=None, init=None, probe_runtime=True):
+ probe_runtime=probe_runtime)
+ self.init_kernel_modules()
+
++ # need to set _host_sysroot before PackageManager()
++ if sysroot:
++ self._container_init()
++ self._host_sysroot = sysroot
++ else:
++ sysroot = self._container_init()
++
+ if init is not None:
+ self.init_system = init
+ elif os.path.isdir("/run/systemd/system/"):
+- self.init_system = SystemdInit()
++ self.init_system = SystemdInit(chroot=sysroot)
+ else:
+ self.init_system = InitSystem()
+
+@@ -130,6 +141,21 @@ def get_local_name(self):
+ def sanitize_filename(self, name):
+ return re.sub(r"[^-a-z,A-Z.0-9]", "", name)
+
++ def _container_init(self):
++ """Check if sos is running in a container and perform container
++ specific initialisation based on ENV_HOST_SYSROOT.
++ """
++ if ENV_CONTAINER in os.environ:
++ if os.environ[ENV_CONTAINER] in ['docker', 'oci', 'podman']:
++ self._in_container = True
++ if ENV_HOST_SYSROOT in os.environ:
++ self._host_sysroot = os.environ[ENV_HOST_SYSROOT]
++ use_sysroot = self._in_container and self._host_sysroot is not None
++ if use_sysroot:
++ host_tmp_dir = os.path.abspath(self._host_sysroot + self._tmp_dir)
++ self._tmp_dir = host_tmp_dir
++ return self._host_sysroot if use_sysroot else None
++
+ def init_kernel_modules(self):
+ """Obtain a list of loaded kernel modules to reference later for plugin
+ enablement and SoSPredicate checks
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index b3a84336be..3476e21fb2 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -17,7 +17,7 @@
+ from sos.presets.redhat import (RHEL_PRESETS, ATOMIC_PRESETS, RHV, RHEL,
+ CB, RHOSP, RHOCP, RH_CFME, RH_SATELLITE,
+ ATOMIC)
+-from sos.policies.distros import LinuxPolicy
++from sos.policies.distros import LinuxPolicy, ENV_HOST_SYSROOT
+ from sos.policies.package_managers.rpm import RpmPackageManager
+ from sos import _sos as _
+
+@@ -56,12 +56,6 @@ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ super(RedHatPolicy, self).__init__(sysroot=sysroot, init=init,
+ probe_runtime=probe_runtime)
+ self.usrmove = False
+- # need to set _host_sysroot before PackageManager()
+- if sysroot:
+- self._container_init()
+- self._host_sysroot = sysroot
+- else:
+- sysroot = self._container_init()
+
+ self.package_manager = RpmPackageManager(chroot=sysroot,
+ remote_exec=remote_exec)
+@@ -140,21 +134,6 @@ def transform_path(path):
+ else:
+ return files
+
+- def _container_init(self):
+- """Check if sos is running in a container and perform container
+- specific initialisation based on ENV_HOST_SYSROOT.
+- """
+- if ENV_CONTAINER in os.environ:
+- if os.environ[ENV_CONTAINER] in ['docker', 'oci', 'podman']:
+- self._in_container = True
+- if ENV_HOST_SYSROOT in os.environ:
+- self._host_sysroot = os.environ[ENV_HOST_SYSROOT]
+- use_sysroot = self._in_container and self._host_sysroot is not None
+- if use_sysroot:
+- host_tmp_dir = os.path.abspath(self._host_sysroot + self._tmp_dir)
+- self._tmp_dir = host_tmp_dir
+- return self._host_sysroot if use_sysroot else None
+-
+ def runlevel_by_service(self, name):
+ from subprocess import Popen, PIPE
+ ret = []
+@@ -183,10 +162,6 @@ def get_tmp_dir(self, opt_tmp_dir):
+ return opt_tmp_dir
+
+
+-# Container environment variables on Red Hat systems.
+-ENV_CONTAINER = 'container'
+-ENV_HOST_SYSROOT = 'HOST'
+-
+ # Legal disclaimer text for Red Hat products
+ disclaimer_text = """
+ Any information provided to %(vendor)s will be treated in \
+diff --git a/sos/policies/init_systems/__init__.py b/sos/policies/init_systems/__init__.py
+index dd663e6522..beac44cee3 100644
+--- a/sos/policies/init_systems/__init__.py
++++ b/sos/policies/init_systems/__init__.py
+@@ -29,9 +29,14 @@ class InitSystem():
+ status of services
+ :type query_cmd: ``str``
+
++ :param chroot: Location to chroot to for any command execution, i.e. the
++ sysroot if we're running in a container
++ :type chroot: ``str`` or ``None``
++
+ """
+
+- def __init__(self, init_cmd=None, list_cmd=None, query_cmd=None):
++ def __init__(self, init_cmd=None, list_cmd=None, query_cmd=None,
++ chroot=None):
+ """Initialize a new InitSystem()"""
+
+ self.services = {}
+@@ -39,6 +44,7 @@ def __init__(self, init_cmd=None, list_cmd=None, query_cmd=None):
+ self.init_cmd = init_cmd
+ self.list_cmd = "%s %s" % (self.init_cmd, list_cmd) or None
+ self.query_cmd = "%s %s" % (self.init_cmd, query_cmd) or None
++ self.chroot = chroot
+
+ def is_enabled(self, name):
+ """Check if given service name is enabled
+@@ -108,7 +114,10 @@ def _query_service(self, name):
+ """Query an individual service"""
+ if self.query_cmd:
+ try:
+- return sos_get_command_output("%s %s" % (self.query_cmd, name))
++ return sos_get_command_output(
++ "%s %s" % (self.query_cmd, name),
++ chroot=self.chroot
++ )
+ except Exception:
+ return None
+ return None
+diff --git a/sos/policies/init_systems/systemd.py b/sos/policies/init_systems/systemd.py
+index 1b138f97b3..76dc57e27f 100644
+--- a/sos/policies/init_systems/systemd.py
++++ b/sos/policies/init_systems/systemd.py
+@@ -15,11 +15,12 @@
+ class SystemdInit(InitSystem):
+ """InitSystem abstraction for SystemD systems"""
+
+- def __init__(self):
++ def __init__(self, chroot=None):
+ super(SystemdInit, self).__init__(
+ init_cmd='systemctl',
+ list_cmd='list-unit-files --type=service',
+- query_cmd='status'
++ query_cmd='status',
++ chroot=chroot
+ )
+ self.load_all_services()
+
+@@ -30,7 +31,7 @@ def parse_query(self, output):
+ return 'unknown'
+
+ def load_all_services(self):
+- svcs = shell_out(self.list_cmd).splitlines()[1:]
++ svcs = shell_out(self.list_cmd, chroot=self.chroot).splitlines()[1:]
+ for line in svcs:
+ try:
+ name = line.split('.service')[0]
+From e869bc84c714bfc2249bbcb84e14908049ee42c4 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 27 Sep 2021 12:07:08 -0400
+Subject: [PATCH] [Plugin,utilities] Add sysroot wrapper for os.path.join
+
+Adds a wrapper for `os.path.join()` which accounts for non-/ sysroots,
+like we have done previously for other `os.path` methods. Further
+updates `Plugin()` to use this wrapper where appropriate.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/__init__.py | 43 +++++++++++++++++-----------------
+ sos/utilities.py | 6 +++++
+ 2 files changed, 28 insertions(+), 21 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index c635b8de9..1f84bca49 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -13,7 +13,7 @@
+ from sos.utilities import (sos_get_command_output, import_module, grep,
+ fileobj, tail, is_executable, TIMEOUT_DEFAULT,
+ path_exists, path_isdir, path_isfile, path_islink,
+- listdir)
++ listdir, path_join)
+
+ import os
+ import glob
+@@ -708,19 +708,6 @@ def _log_info(self, msg):
+ def _log_debug(self, msg):
+ self.soslog.debug(self._format_msg(msg))
+
+- def join_sysroot(self, path):
+- """Join a given path with the configured sysroot
+-
+- :param path: The filesystem path that needs to be joined
+- :type path: ``str``
+-
+- :returns: The joined filesystem path
+- :rtype: ``str``
+- """
+- if path[0] == os.sep:
+- path = path[1:]
+- return os.path.join(self.sysroot, path)
+-
+ def strip_sysroot(self, path):
+ """Remove the configured sysroot from a filesystem path
+
+@@ -1176,7 +1163,7 @@ def _copy_dir(self, srcpath):
+
+ def _get_dest_for_srcpath(self, srcpath):
+ if self.use_sysroot():
+- srcpath = self.join_sysroot(srcpath)
++ srcpath = self.path_join(srcpath)
+ for copied in self.copied_files:
+ if srcpath == copied["srcpath"]:
+ return copied["dstpath"]
+@@ -1284,7 +1271,7 @@ def add_forbidden_path(self, forbidden, recursive=False):
+ forbidden = [forbidden]
+
+ if self.use_sysroot():
+- forbidden = [self.join_sysroot(f) for f in forbidden]
++ forbidden = [self.path_join(f) for f in forbidden]
+
+ for forbid in forbidden:
+ self._log_info("adding forbidden path '%s'" % forbid)
+@@ -1438,7 +1425,7 @@ def add_copy_spec(self, copyspecs, sizelimit=None, maxage=None,
+ since = self.get_option('since')
+
+ logarchive_pattern = re.compile(r'.*((\.(zip|gz|bz2|xz))|[-.][\d]+)$')
+- configfile_pattern = re.compile(r"^%s/*" % self.join_sysroot("etc"))
++ configfile_pattern = re.compile(r"^%s/*" % self.path_join("etc"))
+
+ if not self.test_predicate(pred=pred):
+ self._log_info("skipped copy spec '%s' due to predicate (%s)" %
+@@ -1468,7 +1455,7 @@ def add_copy_spec(self, copyspecs, sizelimit=None, maxage=None,
+ return False
+
+ if self.use_sysroot():
+- copyspec = self.join_sysroot(copyspec)
++ copyspec = self.path_join(copyspec)
+
+ files = self._expand_copy_spec(copyspec)
+
+@@ -1683,7 +1670,7 @@ def _add_device_cmd(self, cmds, devices, timeout=None, sizelimit=None,
+ if not _dev_ok:
+ continue
+ if prepend_path:
+- device = os.path.join(prepend_path, device)
++ device = self.path_join(prepend_path, device)
+ _cmd = cmd % {'dev': device}
+ self._add_cmd_output(cmd=_cmd, timeout=timeout,
+ sizelimit=sizelimit, chroot=chroot,
+@@ -2592,7 +2579,7 @@ def __expand(paths):
+ if self.path_isfile(path) or self.path_islink(path):
+ found_paths.append(path)
+ elif self.path_isdir(path) and self.listdir(path):
+- found_paths.extend(__expand(os.path.join(path, '*')))
++ found_paths.extend(__expand(self.path_join(path, '*')))
+ else:
+ found_paths.append(path)
+ except PermissionError:
+@@ -2608,7 +2595,7 @@ def __expand(paths):
+ if (os.access(copyspec, os.R_OK) and self.path_isdir(copyspec) and
+ self.listdir(copyspec)):
+ # the directory exists and is non-empty, recurse through it
+- copyspec = os.path.join(copyspec, '*')
++ copyspec = self.path_join(copyspec, '*')
+ expanded = glob.glob(copyspec, recursive=True)
+ recursed_files = []
+ for _path in expanded:
+@@ -2877,6 +2864,20 @@ def listdir(self, path):
+ """
+ return listdir(path, self.commons['cmdlineopts'].sysroot)
+
++ def path_join(self, path, *p):
++ """Helper to call the sos.utilities wrapper that allows the
++ corresponding `os` call to account for sysroot
++
++ :param path: The leading path passed to os.path.join()
++ :type path: ``str``
++
++ :param p: Following path section(s) to be joined with ``path``,
++ an empty parameter will result in a path that ends with
++ a separator
++ :type p: ``str``
++ """
++ return path_join(path, *p, sysroot=self.sysroot)
++
+ def postproc(self):
+ """Perform any postprocessing. To be replaced by a plugin if required.
+ """
+diff --git a/sos/utilities.py b/sos/utilities.py
+index c940e066d..b75751539 100644
+--- a/sos/utilities.py
++++ b/sos/utilities.py
+@@ -242,6 +242,12 @@ def listdir(path, sysroot):
+ return _os_wrapper(path, sysroot, 'listdir', os)
+
+
++def path_join(path, *p, sysroot=os.sep):
++ if not path.startswith(sysroot):
++ path = os.path.join(sysroot, path.lstrip(os.sep))
++ return os.path.join(path, *p)
++
++
+ class AsyncReader(threading.Thread):
+ """Used to limit command output to a given size without deadlocking
+ sos.
+From 9596473d1779b9c48e9923c220aaf2b8d9b3bebf Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 18 Nov 2021 13:17:14 -0500
+Subject: [PATCH] [global] Align sysroot determination and usage across sos
+
+The determination of sysroot - being automatic, user-specified, or
+controlled via environment variables in a container - has gotten muddied
+over time. This has resulted in different parts of the project;
+`Policy`, `Plugin`, `SoSComponent`, etc... to not always be in sync when
+sysroot is not `/`, thus causing varying and unexpected/unintended
+behavior.
+
+Fix this by only determining sysroot within `Policy()` initialization,
+and then using that determination across all aspects of the project that
+use or reference sysroot.
+
+This results in several changes:
+
+- `PackageManager()` will now (again) correctly reference host package
+ lists when sos is run in a container.
+
+- `ContainerRuntime()` is now able to activate when sos is running in a
+ container.
+
+- Plugins will now properly use sysroot for _all_ plugin enablement
+ triggers.
+
+- Plugins, Policy, and SoSComponents now all reference the
+ `self.sysroot` variable, rather than changing between `sysroot`.
+`_host_sysroot`, and `commons['sysroot']`. `_host_sysroot` has been
+removed from `Policy`.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/archive.py | 2 +-
+ sos/component.py | 2 +-
+ sos/policies/__init__.py | 11 +----------
+ sos/policies/distros/__init__.py | 33 +++++++++++++++++++------------
+ sos/policies/distros/debian.py | 2 +-
+ sos/policies/distros/redhat.py | 3 +--
+ sos/policies/runtimes/__init__.py | 15 +++++++++-----
+ sos/policies/runtimes/docker.py | 4 ++--
+ sos/report/__init__.py | 6 ++----
+ sos/report/plugins/__init__.py | 22 +++++++++++----------
+ sos/report/plugins/unpackaged.py | 7 ++++---
+ sos/utilities.py | 13 ++++++++----
+ 12 files changed, 64 insertions(+), 56 deletions(-)
+
+diff --git a/sos/archive.py b/sos/archive.py
+index b02b247595..e3c68b7789 100644
+--- a/sos/archive.py
++++ b/sos/archive.py
+@@ -153,7 +153,7 @@ def dest_path(self, name):
+ return (os.path.join(self._archive_root, name))
+
+ def join_sysroot(self, path):
+- if path.startswith(self.sysroot):
++ if not self.sysroot or path.startswith(self.sysroot):
+ return path
+ if path[0] == os.sep:
+ path = path[1:]
+diff --git a/sos/component.py b/sos/component.py
+index 5ac6e47f4f..dba0aabf2b 100644
+--- a/sos/component.py
++++ b/sos/component.py
+@@ -109,7 +109,7 @@ def __init__(self, parser, parsed_args, cmdline_args):
+ try:
+ import sos.policies
+ self.policy = sos.policies.load(sysroot=self.opts.sysroot)
+- self.sysroot = self.policy.host_sysroot()
++ self.sysroot = self.policy.sysroot
+ except KeyboardInterrupt:
+ self._exit(0)
+ self._is_root = self.policy.is_root()
+diff --git a/sos/policies/__init__.py b/sos/policies/__init__.py
+index fb8db1d724..ef9188deb4 100644
+--- a/sos/policies/__init__.py
++++ b/sos/policies/__init__.py
+@@ -110,7 +110,6 @@ class Policy(object):
+ presets = {"": PresetDefaults()}
+ presets_path = PRESETS_PATH
+ _in_container = False
+- _host_sysroot = '/'
+
+ def __init__(self, sysroot=None, probe_runtime=True):
+ """Subclasses that choose to override this initializer should call
+@@ -124,7 +123,7 @@ def __init__(self, sysroot=None, probe_runtime=True):
+ self.package_manager = PackageManager()
+ self.valid_subclasses = [IndependentPlugin]
+ self.set_exec_path()
+- self._host_sysroot = sysroot
++ self.sysroot = sysroot
+ self.register_presets(GENERIC_PRESETS)
+
+ def check(self, remote=''):
+@@ -177,14 +176,6 @@ def in_container(self):
+ """
+ return self._in_container
+
+- def host_sysroot(self):
+- """Get the host's default sysroot
+-
+- :returns: Host sysroot
+- :rtype: ``str`` or ``None``
+- """
+- return self._host_sysroot
+-
+ def dist_version(self):
+ """
+ Return the OS version
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index 7bdc81b852..c69fc1e73c 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -71,19 +71,18 @@ class LinuxPolicy(Policy):
+ def __init__(self, sysroot=None, init=None, probe_runtime=True):
+ super(LinuxPolicy, self).__init__(sysroot=sysroot,
+ probe_runtime=probe_runtime)
+- self.init_kernel_modules()
+
+- # need to set _host_sysroot before PackageManager()
+ if sysroot:
+- self._container_init()
+- self._host_sysroot = sysroot
++ self.sysroot = sysroot
+ else:
+- sysroot = self._container_init()
++ self.sysroot = self._container_init()
++
++ self.init_kernel_modules()
+
+ if init is not None:
+ self.init_system = init
+ elif os.path.isdir("/run/systemd/system/"):
+- self.init_system = SystemdInit(chroot=sysroot)
++ self.init_system = SystemdInit(chroot=self.sysroot)
+ else:
+ self.init_system = InitSystem()
+
+@@ -149,27 +148,30 @@ def _container_init(self):
+ if os.environ[ENV_CONTAINER] in ['docker', 'oci', 'podman']:
+ self._in_container = True
+ if ENV_HOST_SYSROOT in os.environ:
+- self._host_sysroot = os.environ[ENV_HOST_SYSROOT]
+- use_sysroot = self._in_container and self._host_sysroot is not None
++ _host_sysroot = os.environ[ENV_HOST_SYSROOT]
++ use_sysroot = self._in_container and _host_sysroot is not None
+ if use_sysroot:
+- host_tmp_dir = os.path.abspath(self._host_sysroot + self._tmp_dir)
++ host_tmp_dir = os.path.abspath(_host_sysroot + self._tmp_dir)
+ self._tmp_dir = host_tmp_dir
+- return self._host_sysroot if use_sysroot else None
++ return _host_sysroot if use_sysroot else None
+
+ def init_kernel_modules(self):
+ """Obtain a list of loaded kernel modules to reference later for plugin
+ enablement and SoSPredicate checks
+ """
+ self.kernel_mods = []
++ release = os.uname().release
+
+ # first load modules from lsmod
+- lines = shell_out("lsmod", timeout=0).splitlines()
++ lines = shell_out("lsmod", timeout=0, chroot=self.sysroot).splitlines()
+ self.kernel_mods.extend([
+ line.split()[0].strip() for line in lines[1:]
+ ])
+
+ # next, include kernel builtins
+- builtins = "/usr/lib/modules/%s/modules.builtin" % os.uname().release
++ builtins = self.join_sysroot(
++ "/usr/lib/modules/%s/modules.builtin" % release
++ )
+ try:
+ with open(builtins, "r") as mfile:
+ for line in mfile:
+@@ -186,7 +188,7 @@ def init_kernel_modules(self):
+ 'dm_mod': 'CONFIG_BLK_DEV_DM'
+ }
+
+- booted_config = "/boot/config-%s" % os.uname().release
++ booted_config = self.join_sysroot("/boot/config-%s" % release)
+ kconfigs = []
+ try:
+ with open(booted_config, "r") as kfile:
+@@ -200,6 +202,11 @@ def init_kernel_modules(self):
+ if config_strings[builtin] in kconfigs:
+ self.kernel_mods.append(builtin)
+
++ def join_sysroot(self, path):
++ if self.sysroot and self.sysroot != '/':
++ path = os.path.join(self.sysroot, path.lstrip('/'))
++ return path
++
+ def pre_work(self):
+ # this method will be called before the gathering begins
+
+diff --git a/sos/policies/distros/debian.py b/sos/policies/distros/debian.py
+index 95b389a65e..639fd5eba3 100644
+--- a/sos/policies/distros/debian.py
++++ b/sos/policies/distros/debian.py
+@@ -27,7 +27,7 @@ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ remote_exec=None):
+ super(DebianPolicy, self).__init__(sysroot=sysroot, init=init,
+ probe_runtime=probe_runtime)
+- self.package_manager = DpkgPackageManager(chroot=sysroot,
++ self.package_manager = DpkgPackageManager(chroot=self.sysroot,
+ remote_exec=remote_exec)
+ self.valid_subclasses += [DebianPlugin]
+
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index eb44240736..4b14abaf3a 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -42,7 +42,6 @@ class RedHatPolicy(LinuxPolicy):
+ _redhat_release = '/etc/redhat-release'
+ _tmp_dir = "/var/tmp"
+ _in_container = False
+- _host_sysroot = '/'
+ default_scl_prefix = '/opt/rh'
+ name_pattern = 'friendly'
+ upload_url = None
+@@ -57,7 +56,7 @@ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ probe_runtime=probe_runtime)
+ self.usrmove = False
+
+- self.package_manager = RpmPackageManager(chroot=sysroot,
++ self.package_manager = RpmPackageManager(chroot=self.sysroot,
+ remote_exec=remote_exec)
+
+ self.valid_subclasses += [RedHatPlugin]
+diff --git a/sos/policies/runtimes/__init__.py b/sos/policies/runtimes/__init__.py
+index f28d6a1df3..2e60ad2361 100644
+--- a/sos/policies/runtimes/__init__.py
++++ b/sos/policies/runtimes/__init__.py
+@@ -64,7 +64,7 @@ def check_is_active(self):
+ :returns: ``True`` if the runtime is active, else ``False``
+ :rtype: ``bool``
+ """
+- if is_executable(self.binary):
++ if is_executable(self.binary, self.policy.sysroot):
+ self.active = True
+ return True
+ return False
+@@ -78,7 +78,7 @@ def get_containers(self, get_all=False):
+ containers = []
+ _cmd = "%s ps %s" % (self.binary, '-a' if get_all else '')
+ if self.active:
+- out = sos_get_command_output(_cmd)
++ out = sos_get_command_output(_cmd, chroot=self.policy.sysroot)
+ if out['status'] == 0:
+ for ent in out['output'].splitlines()[1:]:
+ ent = ent.split()
+@@ -112,8 +112,10 @@ def get_images(self):
+ images = []
+ fmt = '{{lower .Repository}}:{{lower .Tag}} {{lower .ID}}'
+ if self.active:
+- out = sos_get_command_output("%s images --format '%s'"
+- % (self.binary, fmt))
++ out = sos_get_command_output(
++ "%s images --format '%s'" % (self.binary, fmt),
++ chroot=self.policy.sysroot
++ )
+ if out['status'] == 0:
+ for ent in out['output'].splitlines():
+ ent = ent.split()
+@@ -129,7 +131,10 @@ def get_volumes(self):
+ """
+ vols = []
+ if self.active:
+- out = sos_get_command_output("%s volume ls" % self.binary)
++ out = sos_get_command_output(
++ "%s volume ls" % self.binary,
++ chroot=self.policy.sysroot
++ )
+ if out['status'] == 0:
+ for ent in out['output'].splitlines()[1:]:
+ ent = ent.split()
+diff --git a/sos/policies/runtimes/docker.py b/sos/policies/runtimes/docker.py
+index 759dfaf6a0..e81f580ec3 100644
+--- a/sos/policies/runtimes/docker.py
++++ b/sos/policies/runtimes/docker.py
+@@ -18,9 +18,9 @@ class DockerContainerRuntime(ContainerRuntime):
+ name = 'docker'
+ binary = 'docker'
+
+- def check_is_active(self):
++ def check_is_active(self, sysroot=None):
+ # the daemon must be running
+- if (is_executable('docker') and
++ if (is_executable('docker', sysroot) and
+ (self.policy.init_system.is_running('docker') or
+ self.policy.init_system.is_running('snap.docker.dockerd'))):
+ self.active = True
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index a4c92accd3..a6c72778fc 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -173,14 +173,12 @@ def __init__(self, parser, args, cmdline):
+ self._set_directories()
+
+ msg = "default"
+- host_sysroot = self.policy.host_sysroot()
++ self.sysroot = self.policy.sysroot
+ # set alternate system root directory
+ if self.opts.sysroot:
+ msg = "cmdline"
+- self.sysroot = self.opts.sysroot
+- elif self.policy.in_container() and host_sysroot != os.sep:
++ elif self.policy.in_container() and self.sysroot != os.sep:
+ msg = "policy"
+- self.sysroot = host_sysroot
+ self.soslog.debug("set sysroot to '%s' (%s)" % (self.sysroot, msg))
+
+ if self.opts.chroot not in chroot_modes:
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 46028bb124..e180ae1727 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -724,7 +724,7 @@ def strip_sysroot(self, path):
+ """
+ if not self.use_sysroot():
+ return path
+- if path.startswith(self.sysroot):
++ if self.sysroot and path.startswith(self.sysroot):
+ return path[len(self.sysroot):]
+ return path
+
+@@ -743,8 +743,10 @@ def tmp_in_sysroot(self):
+ ``False``
+ :rtype: ``bool``
+ """
+- paths = [self.sysroot, self.archive.get_tmp_dir()]
+- return os.path.commonprefix(paths) == self.sysroot
++ # if sysroot is still None, that implies '/'
++ _sysroot = self.sysroot or '/'
++ paths = [_sysroot, self.archive.get_tmp_dir()]
++ return os.path.commonprefix(paths) == _sysroot
+
+ def is_installed(self, package_name):
+ """Is the package $package_name installed?
+@@ -2621,7 +2623,7 @@ def __expand(paths):
+ return list(set(expanded))
+
+ def _collect_copy_specs(self):
+- for path in self.copy_paths:
++ for path in sorted(self.copy_paths, reverse=True):
+ self._log_info("collecting path '%s'" % path)
+ self._do_copy_path(path)
+ self.generate_copyspec_tags()
+@@ -2749,7 +2751,7 @@ def _check_plugin_triggers(self, files, packages, commands, services,
+
+ return ((any(self.path_exists(fname) for fname in files) or
+ any(self.is_installed(pkg) for pkg in packages) or
+- any(is_executable(cmd) for cmd in commands) or
++ any(is_executable(cmd, self.sysroot) for cmd in commands) or
+ any(self.is_module_loaded(mod) for mod in self.kernel_mods) or
+ any(self.is_service(svc) for svc in services) or
+ any(self.container_exists(cntr) for cntr in containers)) and
+@@ -2817,7 +2819,7 @@ def path_exists(self, path):
+ :returns: True if the path exists in sysroot, else False
+ :rtype: ``bool``
+ """
+- return path_exists(path, self.commons['cmdlineopts'].sysroot)
++ return path_exists(path, self.sysroot)
+
+ def path_isdir(self, path):
+ """Helper to call the sos.utilities wrapper that allows the
+@@ -2830,7 +2832,7 @@ def path_isdir(self, path):
+ :returns: True if the path is a dir, else False
+ :rtype: ``bool``
+ """
+- return path_isdir(path, self.commons['cmdlineopts'].sysroot)
++ return path_isdir(path, self.sysroot)
+
+ def path_isfile(self, path):
+ """Helper to call the sos.utilities wrapper that allows the
+@@ -2843,7 +2845,7 @@ def path_isfile(self, path):
+ :returns: True if the path is a file, else False
+ :rtype: ``bool``
+ """
+- return path_isfile(path, self.commons['cmdlineopts'].sysroot)
++ return path_isfile(path, self.sysroot)
+
+ def path_islink(self, path):
+ """Helper to call the sos.utilities wrapper that allows the
+@@ -2856,7 +2858,7 @@ def path_islink(self, path):
+ :returns: True if the path is a link, else False
+ :rtype: ``bool``
+ """
+- return path_islink(path, self.commons['cmdlineopts'].sysroot)
++ return path_islink(path, self.sysroot)
+
+ def listdir(self, path):
+ """Helper to call the sos.utilities wrapper that allows the
+@@ -2869,7 +2871,7 @@ def listdir(self, path):
+ :returns: Contents of path, if it is a directory
+ :rtype: ``list``
+ """
+- return listdir(path, self.commons['cmdlineopts'].sysroot)
++ return listdir(path, self.sysroot)
+
+ def path_join(self, path, *p):
+ """Helper to call the sos.utilities wrapper that allows the
+diff --git a/sos/report/plugins/unpackaged.py b/sos/report/plugins/unpackaged.py
+index 772b1d1fbb..24203c4b13 100644
+--- a/sos/report/plugins/unpackaged.py
++++ b/sos/report/plugins/unpackaged.py
+@@ -58,10 +58,11 @@ def format_output(files):
+ """
+ expanded = []
+ for f in files:
+- if self.path_islink(f):
+- expanded.append("{} -> {}".format(f, os.readlink(f)))
++ fp = self.path_join(f)
++ if self.path_islink(fp):
++ expanded.append("{} -> {}".format(fp, os.readlink(fp)))
+ else:
+- expanded.append(f)
++ expanded.append(fp)
+ return expanded
+
+ # Check command predicate to avoid costly processing
+diff --git a/sos/utilities.py b/sos/utilities.py
+index b757515397..d66309334b 100644
+--- a/sos/utilities.py
++++ b/sos/utilities.py
+@@ -96,11 +96,15 @@ def grep(pattern, *files_or_paths):
+ return matches
+
+
+-def is_executable(command):
++def is_executable(command, sysroot=None):
+ """Returns if a command matches an executable on the PATH"""
+
+ paths = os.environ.get("PATH", "").split(os.path.pathsep)
+ candidates = [command] + [os.path.join(p, command) for p in paths]
++ if sysroot:
++ candidates += [
++ os.path.join(sysroot, c.lstrip('/')) for c in candidates
++ ]
+ return any(os.access(path, os.X_OK) for path in candidates)
+
+
+@@ -216,8 +220,9 @@ def get_human_readable(size, precision=2):
+
+
+ def _os_wrapper(path, sysroot, method, module=os.path):
+- if sysroot not in [None, '/']:
+- path = os.path.join(sysroot, path.lstrip('/'))
++ if sysroot and sysroot != os.sep:
++ if not path.startswith(sysroot):
++ path = os.path.join(sysroot, path.lstrip('/'))
+ _meth = getattr(module, method)
+ return _meth(path)
+
+@@ -243,7 +248,7 @@ def listdir(path, sysroot):
+
+
+ def path_join(path, *p, sysroot=os.sep):
+- if not path.startswith(sysroot):
++ if sysroot and not path.startswith(sysroot):
+ path = os.path.join(sysroot, path.lstrip(os.sep))
+ return os.path.join(path, *p)
+
+From a43124e1f6217107838eed4d70339d100cbbc77a Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 9 Feb 2022 19:45:27 +0100
+Subject: [PATCH] [policies] Set fallback to None sysroot
+
+9596473 commit added a regression allowing to set sysroot to None
+when running sos report on a regular system (outside a container). In
+such a case, we need to fallback to '/' sysroot.
+
+Resolves: #2846
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/__init__.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index f3c1de11..9048f1c4 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -78,7 +78,7 @@ class LinuxPolicy(Policy):
+ if sysroot:
+ self.sysroot = sysroot
+ else:
+- self.sysroot = self._container_init()
++ self.sysroot = self._container_init() or '/'
+
+ self.init_kernel_modules()
+
+--
+2.34.1
+
diff --git a/SOURCES/sos-bz2011538-iptables-save-under-nf_tables-kmod.patch b/SOURCES/sos-bz2011538-iptables-save-under-nf_tables-kmod.patch
new file mode 100644
index 0000000..e234bc6
--- /dev/null
+++ b/SOURCES/sos-bz2011538-iptables-save-under-nf_tables-kmod.patch
@@ -0,0 +1,73 @@
+From 7d5157aa5071e3620246e2d4aa80acb2d3ed30f0 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Tue, 28 Sep 2021 22:44:52 +0200
+Subject: [PATCH] [networking] prevent iptables-save commands to load nf_tables
+ kmod
+
+If iptables has built-in nf_tables kmod, then
+'ip netns <foo> iptables-save' command requires the kmod which must
+be guarded by predicate.
+
+Analogously for ip6tables.
+
+Resolves: #2703
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/networking.py | 29 ++++++++++++++++++++++++-----
+ 1 file changed, 24 insertions(+), 5 deletions(-)
+
+diff --git a/sos/report/plugins/networking.py b/sos/report/plugins/networking.py
+index c80ae719..1237f629 100644
+--- a/sos/report/plugins/networking.py
++++ b/sos/report/plugins/networking.py
+@@ -182,22 +182,41 @@ class Networking(Plugin):
+ # per-namespace.
+ self.add_cmd_output("ip netns")
+ cmd_prefix = "ip netns exec "
+- for namespace in self.get_network_namespaces(
+- self.get_option("namespace_pattern"),
+- self.get_option("namespaces")):
++ namespaces = self.get_network_namespaces(
++ self.get_option("namespace_pattern"),
++ self.get_option("namespaces"))
++ if (namespaces):
++ # 'ip netns exec <foo> iptables-save' must be guarded by nf_tables
++ # kmod, if 'iptables -V' output contains 'nf_tables'
++ # analogously for ip6tables
++ co = {'cmd': 'iptables -V', 'output': 'nf_tables'}
++ co6 = {'cmd': 'ip6tables -V', 'output': 'nf_tables'}
++ iptables_with_nft = (SoSPredicate(self, kmods=['nf_tables'])
++ if self.test_predicate(self,
++ pred=SoSPredicate(self, cmd_outputs=co))
++ else None)
++ ip6tables_with_nft = (SoSPredicate(self, kmods=['nf_tables'])
++ if self.test_predicate(self,
++ pred=SoSPredicate(self, cmd_outputs=co6))
++ else None)
++ for namespace in namespaces:
+ ns_cmd_prefix = cmd_prefix + namespace + " "
+ self.add_cmd_output([
+ ns_cmd_prefix + "ip address show",
+ ns_cmd_prefix + "ip route show table all",
+ ns_cmd_prefix + "ip -s -s neigh show",
+ ns_cmd_prefix + "ip rule list",
+- ns_cmd_prefix + "iptables-save",
+- ns_cmd_prefix + "ip6tables-save",
+ ns_cmd_prefix + "netstat %s -neopa" % self.ns_wide,
+ ns_cmd_prefix + "netstat -s",
+ ns_cmd_prefix + "netstat %s -agn" % self.ns_wide,
+ ns_cmd_prefix + "nstat -zas",
+ ], priority=50)
++ self.add_cmd_output([ns_cmd_prefix + "iptables-save"],
++ pred=iptables_with_nft,
++ priority=50)
++ self.add_cmd_output([ns_cmd_prefix + "ip6tables-save"],
++ pred=ip6tables_with_nft,
++ priority=50)
+
+ ss_cmd = ns_cmd_prefix + "ss -peaonmi"
+ # --allow-system-changes is handled directly in predicate
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2012858-dryrun-uncaught-exception.patch b/SOURCES/sos-bz2012858-dryrun-uncaught-exception.patch
new file mode 100644
index 0000000..619d538
--- /dev/null
+++ b/SOURCES/sos-bz2012858-dryrun-uncaught-exception.patch
@@ -0,0 +1,33 @@
+From e56b3ea999731b831ebba80cf367e43e65c12b62 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 4 Oct 2021 14:43:08 +0200
+Subject: [PATCH] [report] Overwrite pred=None before refering predicate
+ attributes
+
+During a dry run, add_journal method sets pred=None whilst log_skipped_cmd
+refers to predicate attributes. In that case, replace None predicate
+by a default / empty predicate.
+
+Resolves: #2711
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/__init__.py | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 3c2b64d9..c635b8de 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -1693,6 +1693,8 @@ class Plugin():
+ def _add_cmd_output(self, **kwargs):
+ """Internal helper to add a single command to the collection list."""
+ pred = kwargs.pop('pred') if 'pred' in kwargs else SoSPredicate(self)
++ if pred is None:
++ pred = SoSPredicate(self)
+ if 'priority' not in kwargs:
+ kwargs['priority'] = 10
+ if 'changes' not in kwargs:
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2012859-plugin-timeout-unhandled-exception.patch b/SOURCES/sos-bz2012859-plugin-timeout-unhandled-exception.patch
new file mode 100644
index 0000000..e977fb5
--- /dev/null
+++ b/SOURCES/sos-bz2012859-plugin-timeout-unhandled-exception.patch
@@ -0,0 +1,31 @@
+From a93e118a9c88df52fd2c701d2276185f877d565c Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 3 Nov 2021 16:07:15 +0100
+Subject: [PATCH] [report] shutdown threads for timeouted plugins
+
+Wait for shutting down threads of timeouted plugins, to prevent
+them in writing to moved auxiliary files like sos_logs/sos.log
+
+Resolves: #2722
+Closes: #2746
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/__init__.py | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index 1b5bc97d..ef86b28d 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -1046,6 +1046,7 @@ class SoSReport(SoSComponent):
+ self.ui_log.error("\n Plugin %s timed out\n" % plugin[1])
+ self.running_plugs.remove(plugin[1])
+ self.loaded_plugins[plugin[0]-1][1].set_timeout_hit()
++ pool.shutdown(wait=True)
+ pool._threads.clear()
+ if self.opts.estimate_only:
+ from pathlib import Path
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2019697-openvswitch-offline-analysis.patch b/SOURCES/sos-bz2019697-openvswitch-offline-analysis.patch
new file mode 100644
index 0000000..8bd4adb
--- /dev/null
+++ b/SOURCES/sos-bz2019697-openvswitch-offline-analysis.patch
@@ -0,0 +1,151 @@
+From 3f0ec3e55e7dcec89dd7fad10084ea7f16178608 Mon Sep 17 00:00:00 2001
+From: Salvatore Daniele <sdaniele@redhat.com>
+Date: Tue, 7 Sep 2021 13:48:22 -0400
+Subject: [PATCH 1/2] [openvswitch] add ovs default OpenFlow protocols
+
+ovs-vsctl list bridge can return an empty 'protocol' column even when
+there are OpenFlow protocols in place by default.
+
+ovs-ofctl --version will return the range of supported ofp and should
+also be used to ensure flow information for relevant protocol versions
+is collected.
+
+OpenFlow default versions:
+https://docs.openvswitch.org/en/latest/faq/openflow/
+
+Signed-off-by: Salvatore Daniele <sdaniele@redhat.com>
+---
+ sos/report/plugins/openvswitch.py | 26 ++++++++++++++++++++++++++
+ 1 file changed, 26 insertions(+)
+
+diff --git a/sos/report/plugins/openvswitch.py b/sos/report/plugins/openvswitch.py
+index cd897db2..92cc7259 100644
+--- a/sos/report/plugins/openvswitch.py
++++ b/sos/report/plugins/openvswitch.py
+@@ -206,6 +206,7 @@ class OpenVSwitch(Plugin):
+
+ # Gather additional output for each OVS bridge on the host.
+ br_list_result = self.collect_cmd_output("ovs-vsctl -t 5 list-br")
++ ofp_ver_result = self.collect_cmd_output("ovs-ofctl -t 5 --version")
+ if br_list_result['status'] == 0:
+ for br in br_list_result['output'].splitlines():
+ self.add_cmd_output([
+@@ -232,6 +233,16 @@ class OpenVSwitch(Plugin):
+ "OpenFlow15"
+ ]
+
++ # Flow protocol hex identifiers
++ ofp_versions = {
++ 0x01: "OpenFlow10",
++ 0x02: "OpenFlow11",
++ 0x03: "OpenFlow12",
++ 0x04: "OpenFlow13",
++ 0x05: "OpenFlow14",
++ 0x06: "OpenFlow15",
++ }
++
+ # List protocols currently in use, if any
+ ovs_list_bridge_cmd = "ovs-vsctl -t 5 list bridge %s" % br
+ br_info = self.collect_cmd_output(ovs_list_bridge_cmd)
+@@ -242,6 +253,21 @@ class OpenVSwitch(Plugin):
+ br_protos_ln = line[line.find("[")+1:line.find("]")]
+ br_protos = br_protos_ln.replace('"', '').split(", ")
+
++ # If 'list bridge' yeilded no protocols, use the range of
++ # protocols enabled by default on this version of ovs.
++ if br_protos == [''] and ofp_ver_result['output']:
++ ofp_version_range = ofp_ver_result['output'].splitlines()
++ ver_range = []
++
++ for line in ofp_version_range:
++ if "OpenFlow versions" in line:
++ v = line.split("OpenFlow versions ")[1].split(":")
++ ver_range = range(int(v[0], 16), int(v[1], 16)+1)
++
++ for protocol in ver_range:
++ if protocol in ofp_versions:
++ br_protos.append(ofp_versions[protocol])
++
+ # Collect flow information for relevant protocol versions only
+ for flow in flow_versions:
+ if flow in br_protos:
+--
+2.31.1
+
+
+From 5a006024f730213a726c70e82c5ecd2daf685b2b Mon Sep 17 00:00:00 2001
+From: Salvatore Daniele <sdaniele@redhat.com>
+Date: Tue, 7 Sep 2021 14:17:19 -0400
+Subject: [PATCH 2/2] [openvswitch] add commands for offline analysis
+
+Replicas of ovs-vswitchd and ovsdb-server can be recreated offline
+using flow, group, and tlv dumps, and ovs conf.db. This allows for
+offline anaylsis and the use of tools such as ovs-appctl
+ofproto/trace and ovs-ofctl for debugging.
+
+This patch ensures this information is available in the sos report.
+The db is copied rather than collected using ovsdb-client list dump
+for two reasons:
+
+ovsdb-client requires interacting with the ovsdb-server which could
+take it 'down' for some time, and impact large, busy clusters.
+
+The list-dump is not in a format that can be used to restore the db
+offline. All of the information in the list dump is available and more
+by copying the db.
+
+Signed-off-by: Salvatore Daniele <sdaniele@redhat.com>
+---
+ sos/report/plugins/openvswitch.py | 12 ++++++++++--
+ sos/report/plugins/ovn_central.py | 1 +
+ 2 files changed, 11 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/openvswitch.py b/sos/report/plugins/openvswitch.py
+index 92cc7259..003596c6 100644
+--- a/sos/report/plugins/openvswitch.py
++++ b/sos/report/plugins/openvswitch.py
+@@ -75,12 +75,19 @@ class OpenVSwitch(Plugin):
+ "/run/openvswitch/ovs-monitor-ipsec.pid"
+ ])
+
++ self.add_copy_spec([
++ path_join('/usr/local/etc/openvswitch', 'conf.db'),
++ path_join('/etc/openvswitch', 'conf.db'),
++ path_join('/var/lib/openvswitch', 'conf.db'),
++ ])
++ ovs_dbdir = environ.get('OVS_DBDIR')
++ if ovs_dbdir:
++ self.add_copy_spec(path_join(ovs_dbdir, 'conf.db'))
++
+ self.add_cmd_output([
+ # The '-t 5' adds an upper bound on how long to wait to connect
+ # to the Open vSwitch server, avoiding hangs when running sos.
+ "ovs-vsctl -t 5 show",
+- # Gather the database.
+- "ovsdb-client -f list dump",
+ # List the contents of important runtime directories
+ "ls -laZ /run/openvswitch",
+ "ls -laZ /dev/hugepages/",
+@@ -276,6 +283,7 @@ class OpenVSwitch(Plugin):
+ "ovs-ofctl -O %s dump-groups %s" % (flow, br),
+ "ovs-ofctl -O %s dump-group-stats %s" % (flow, br),
+ "ovs-ofctl -O %s dump-flows %s" % (flow, br),
++ "ovs-ofctl -O %s dump-tlv-map %s" % (flow, br),
+ "ovs-ofctl -O %s dump-ports-desc %s" % (flow, br)
+ ])
+
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index a4c483a9..d6647aad 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -138,6 +138,7 @@ class OVNCentral(Plugin):
+ os.path.join('/usr/local/etc/openvswitch', dbfile),
+ os.path.join('/etc/openvswitch', dbfile),
+ os.path.join('/var/lib/openvswitch', dbfile),
++ os.path.join('/var/lib/ovn/etc', dbfile),
+ ])
+ if ovs_dbdir:
+ self.add_copy_spec(os.path.join(ovs_dbdir, dbfile))
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2020778-filter-namespace-per-pattern.patch b/SOURCES/sos-bz2020778-filter-namespace-per-pattern.patch
new file mode 100644
index 0000000..5b0afdb
--- /dev/null
+++ b/SOURCES/sos-bz2020778-filter-namespace-per-pattern.patch
@@ -0,0 +1,54 @@
+From 568eb2fbcf74ecad00d5c06989f55f8a6a9e3516 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Thu, 4 Nov 2021 23:14:21 +0100
+Subject: [PATCH] [report] fix filter_namespace per pattern
+
+Curently, -k networking.namespace_pattern=.. is broken as the R.E. test
+forgets to add the namespace in case of positive match.
+
+Also ensure both plugopts namespace_pattern and namespaces work
+together.
+
+Resolves: #2748
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/__init__.py | 15 +++++++--------
+ 1 file changed, 7 insertions(+), 8 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 3e717993..a0d4e95d 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -2953,21 +2953,20 @@ class Plugin():
+ )
+ for ns in ns_list:
+ # if ns_pattern defined, skip namespaces not matching the pattern
+- if ns_pattern:
+- if not bool(re.match(pattern, ns)):
+- continue
++ if ns_pattern and not bool(re.match(pattern, ns)):
++ continue
++ out_ns.append(ns)
+
+- # if ns_max is defined at all, limit returned list to that number
++ # if ns_max is defined at all, break the loop when the limit is
++ # reached
+ # this allows the use of both '0' and `None` to mean unlimited
+- elif ns_max:
+- out_ns.append(ns)
++ if ns_max:
+ if len(out_ns) == ns_max:
+ self._log_warn("Limiting namespace iteration "
+ "to first %s namespaces found"
+ % ns_max)
+ break
+- else:
+- out_ns.append(ns)
++
+ return out_ns
+
+
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2023481-plugin-timeouts-proper-handling.patch b/SOURCES/sos-bz2023481-plugin-timeouts-proper-handling.patch
new file mode 100644
index 0000000..9fc7c3d
--- /dev/null
+++ b/SOURCES/sos-bz2023481-plugin-timeouts-proper-handling.patch
@@ -0,0 +1,91 @@
+From 3fea9a564c4112d04f6324df0d8b212e78feb5b3 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 3 Nov 2021 11:02:54 -0400
+Subject: [PATCH] [Plugin] Ensure specific plugin timeouts are only set for
+ that plugin
+
+It was discovered that setting a specific plugin timeout via the `-k
+$plugin.timeout` option could influence the timeout setting for other
+plugins that are not also having their timeout explicitly set. Fix this
+by moving the default plugin opts into `Plugin.__init__()` so that each
+plugin is ensured a private copy of these default plugin options.
+
+Additionally, add more timeout data to plugin manifest entries to allow
+for better tracking of this setting.
+
+Adds a test case for this scenario.
+
+Closes: #2744
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/__init__.py | 2 +-
+ sos/report/plugins/__init__.py | 28 +++++++++++++------
+ tests/vendor_tests/redhat/rhbz2018033.py | 35 ++++++++++++++++++++++++
+ 3 files changed, 55 insertions(+), 10 deletions(-)
+ create mode 100644 tests/vendor_tests/redhat/rhbz2018033.py
+
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index ef86b28d..c95e6300 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -766,7 +766,7 @@ class SoSReport(SoSComponent):
+ if self.all_options:
+ self.ui_log.info(_("The following options are available for ALL "
+ "plugins:"))
+- for opt in self.all_options[0][0]._default_plug_opts:
++ for opt in self.all_options[0][0].get_default_plugin_opts():
+ val = opt[3]
+ if val == -1:
+ val = TIMEOUT_DEFAULT
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 49f1af27..3e717993 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -474,12 +474,6 @@ class Plugin(object):
+ # Default predicates
+ predicate = None
+ cmd_predicate = None
+- _default_plug_opts = [
+- ('timeout', 'Timeout in seconds for plugin to finish', 'fast', -1),
+- ('cmd-timeout', 'Timeout in seconds for a command', 'fast', -1),
+- ('postproc', 'Enable post-processing collected plugin data', 'fast',
+- True)
+- ]
+
+ def __init__(self, commons):
+
+@@ -506,7 +500,7 @@ class Plugin(object):
+ else logging.getLogger('sos')
+
+ # add the default plugin opts
+- self.option_list.extend(self._default_plug_opts)
++ self.option_list.extend(self.get_default_plugin_opts())
+
+ # get the option list into a dictionary
+ for opt in self.option_list:
+@@ -591,6 +583,14 @@ class Plugin():
+ # Initialise the default --dry-run predicate
+ self.set_predicate(SoSPredicate(self))
+
++ def get_default_plugin_opts(self):
++ return [
++ ('timeout', 'Timeout in seconds for plugin to finish', 'fast', -1),
++ ('cmd-timeout', 'Timeout in seconds for a command', 'fast', -1),
++ ('postproc', 'Enable post-processing collected plugin data', 'fast',
++ True)
++ ]
++
+ def set_plugin_manifest(self, manifest):
+ """Pass in a manifest object to the plugin to write to
+
+@@ -547,7 +541,9 @@ class Plugin(object):
+ self.manifest.add_field('setup_start', '')
+ self.manifest.add_field('setup_end', '')
+ self.manifest.add_field('setup_time', '')
++ self.manifest.add_field('timeout', self.timeout)
+ self.manifest.add_field('timeout_hit', False)
++ self.manifest.add_field('command_timeout', self.cmdtimeout)
+ self.manifest.add_list('commands', [])
+ self.manifest.add_list('files', [])
+
diff --git a/SOURCES/sos-bz2024893-cleaner-hostnames-improvements.patch b/SOURCES/sos-bz2024893-cleaner-hostnames-improvements.patch
new file mode 100644
index 0000000..b129f9e
--- /dev/null
+++ b/SOURCES/sos-bz2024893-cleaner-hostnames-improvements.patch
@@ -0,0 +1,1829 @@
+From decd39b7799a0579ea085b0da0728b6eabd49b38 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 1 Sep 2021 00:28:58 -0400
+Subject: [PATCH] [clean] Provide archive abstractions to obfuscate more than
+ sos archives
+
+This commit removes the restriction imposed on `sos clean` since its
+introduction in sos-4.0 to only work against known sos report archives
+or build directories. This is because there has been interest in using
+the obfuscation bits of sos in other data-collector projects.
+
+The `SoSObfuscationArchive()` class has been revamped to now be an
+abstraction for different types of archives, and the cleaner logic has
+been updated to leverage this new abstraction rather than assuming we're
+working on an sos archive.
+
+Abstractions are added for our own native use cases - that being `sos
+report` and `sos collect` for at-runtime obfuscation, as well as
+standalone archives previously generated. Further generic abstractions
+are available for plain directories and tarballs however these will not
+provide the same level of coverage as fully supported archive types, as
+is noted in the manpage for sos-clean.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ man/en/sos-clean.1 | 25 ++
+ sos/cleaner/__init__.py | 308 +++++++++---------
+ .../__init__.py} | 80 ++++-
+ sos/cleaner/archives/generic.py | 52 +++
+ sos/cleaner/archives/sos.py | 106 ++++++
+ sos/cleaner/parsers/__init__.py | 6 -
+ sos/cleaner/parsers/hostname_parser.py | 1 -
+ sos/cleaner/parsers/ip_parser.py | 1 -
+ sos/cleaner/parsers/keyword_parser.py | 1 -
+ sos/cleaner/parsers/mac_parser.py | 1 -
+ sos/cleaner/parsers/username_parser.py | 8 -
+ tests/cleaner_tests/existing_archive.py | 7 +
+ tests/cleaner_tests/full_report_run.py | 3 +
+ tests/cleaner_tests/report_with_mask.py | 3 +
+ 14 files changed, 423 insertions(+), 179 deletions(-)
+ rename sos/cleaner/{obfuscation_archive.py => archives/__init__.py} (81%)
+ create mode 100644 sos/cleaner/archives/generic.py
+ create mode 100644 sos/cleaner/archives/sos.py
+
+diff --git a/man/en/sos-clean.1 b/man/en/sos-clean.1
+index b77bc63c..54026713 100644
+--- a/man/en/sos-clean.1
++++ b/man/en/sos-clean.1
+@@ -10,6 +10,7 @@ sos clean - Obfuscate sensitive data from one or more sosreports
+ [\-\-jobs]
+ [\-\-no-update]
+ [\-\-keep-binary-files]
++ [\-\-archive-type]
+
+ .SH DESCRIPTION
+ \fBsos clean\fR or \fBsos mask\fR is an sos subcommand used to obfuscate sensitive information from
+@@ -88,6 +89,30 @@ Users should review any archive that keeps binary files in place before sending
+ a third party.
+
+ Default: False (remove encountered binary files)
++.TP
++.B \-\-archive-type TYPE
++Specify the type of archive that TARGET was generated as.
++When sos inspects a TARGET archive, it tries to identify what type of archive it is.
++For example, it may be a report generated by \fBsos report\fR, or a collection of those
++reports generated by \fBsos collect\fR, which require separate approaches.
++
++This option may be useful if a given TARGET archive is known to be of a specific type,
++but due to unknown reasons or some malformed/missing information in the archive directly,
++that is not properly identified by sos.
++
++The following are accepted values for this option:
++
++ \fBauto\fR Automatically detect the archive type
++ \fBreport\fR An archive generated by \fBsos report\fR
++ \fBcollect\fR An archive generated by \fBsos collect\fR
++
++The following may also be used, however note that these do not attempt to pre-load
++any information from the archives into the parsers. This means that, among other limitations,
++items like host and domain names may not be obfuscated unless an obfuscated mapping already exists
++on the system from a previous execution.
++
++ \fBdata-dir\fR A plain directory on the filesystem.
++ \fBtarball\fR A generic tar archive not associated with any known tool
+
+ .SH SEE ALSO
+ .BR sos (1)
+diff --git a/sos/cleaner/__init__.py b/sos/cleaner/__init__.py
+index 6aadfe79..6d2eb483 100644
+--- a/sos/cleaner/__init__.py
++++ b/sos/cleaner/__init__.py
+@@ -12,9 +12,7 @@ import hashlib
+ import json
+ import logging
+ import os
+-import re
+ import shutil
+-import tarfile
+ import tempfile
+
+ from concurrent.futures import ThreadPoolExecutor
+@@ -27,7 +25,10 @@ from sos.cleaner.parsers.mac_parser import SoSMacParser
+ from sos.cleaner.parsers.hostname_parser import SoSHostnameParser
+ from sos.cleaner.parsers.keyword_parser import SoSKeywordParser
+ from sos.cleaner.parsers.username_parser import SoSUsernameParser
+-from sos.cleaner.obfuscation_archive import SoSObfuscationArchive
++from sos.cleaner.archives.sos import (SoSReportArchive, SoSReportDirectory,
++ SoSCollectorArchive,
++ SoSCollectorDirectory)
++from sos.cleaner.archives.generic import DataDirArchive, TarballArchive
+ from sos.utilities import get_human_readable
+ from textwrap import fill
+
+@@ -41,6 +42,7 @@ class SoSCleaner(SoSComponent):
+ desc = "Obfuscate sensitive networking information in a report"
+
+ arg_defaults = {
++ 'archive_type': 'auto',
+ 'domains': [],
+ 'jobs': 4,
+ 'keywords': [],
+@@ -70,6 +72,7 @@ class SoSCleaner(SoSComponent):
+ self.from_cmdline = False
+ if not hasattr(self.opts, 'jobs'):
+ self.opts.jobs = 4
++ self.opts.archive_type = 'auto'
+ self.soslog = logging.getLogger('sos')
+ self.ui_log = logging.getLogger('sos_ui')
+ # create the tmp subdir here to avoid a potential race condition
+@@ -92,6 +95,17 @@ class SoSCleaner(SoSComponent):
+ SoSUsernameParser(self.cleaner_mapping, self.opts.usernames)
+ ]
+
++ self.archive_types = [
++ SoSReportDirectory,
++ SoSReportArchive,
++ SoSCollectorDirectory,
++ SoSCollectorArchive,
++ # make sure these two are always last as they are fallbacks
++ DataDirArchive,
++ TarballArchive
++ ]
++ self.nested_archive = None
++
+ self.log_info("Cleaner initialized. From cmdline: %s"
+ % self.from_cmdline)
+
+@@ -178,6 +192,11 @@ third party.
+ )
+ clean_grp.add_argument('target', metavar='TARGET',
+ help='The directory or archive to obfuscate')
++ clean_grp.add_argument('--archive-type', default='auto',
++ choices=['auto', 'report', 'collect',
++ 'data-dir', 'tarball'],
++ help=('Specify what kind of archive the target '
++ 'was generated as'))
+ clean_grp.add_argument('--domains', action='extend', default=[],
+ help='List of domain names to obfuscate')
+ clean_grp.add_argument('-j', '--jobs', default=4, type=int,
+@@ -218,59 +237,28 @@ third party.
+
+ In the event the target path is not an archive, abort.
+ """
+- if not tarfile.is_tarfile(self.opts.target):
+- self.ui_log.error(
+- "Invalid target: must be directory or tar archive"
+- )
+- self._exit(1)
+-
+- archive = tarfile.open(self.opts.target)
+- self.arc_name = self.opts.target.split('/')[-1].split('.')[:-2][0]
+-
+- try:
+- archive.getmember(os.path.join(self.arc_name, 'sos_logs'))
+- except Exception:
+- # this is not an sos archive
+- self.ui_log.error("Invalid target: not an sos archive")
+- self._exit(1)
+-
+- # see if there are archives within this archive
+- nested_archives = []
+- for _file in archive.getmembers():
+- if (re.match('sosreport-.*.tar', _file.name.split('/')[-1]) and not
+- (_file.name.endswith(('.md5', '.sha256')))):
+- nested_archives.append(_file.name.split('/')[-1])
+-
+- if nested_archives:
+- self.log_info("Found nested archive(s), extracting top level")
+- nested_path = self.extract_archive(archive)
+- for arc_file in os.listdir(nested_path):
+- if re.match('sosreport.*.tar.*', arc_file):
+- if arc_file.endswith(('.md5', '.sha256')):
+- continue
+- self.report_paths.append(os.path.join(nested_path,
+- arc_file))
+- # add the toplevel extracted archive
+- self.report_paths.append(nested_path)
++ _arc = None
++ if self.opts.archive_type != 'auto':
++ check_type = self.opts.archive_type.replace('-', '_')
++ for archive in self.archive_types:
++ if archive.type_name == check_type:
++ _arc = archive(self.opts.target, self.tmpdir)
+ else:
+- self.report_paths.append(self.opts.target)
+-
+- archive.close()
+-
+- def extract_archive(self, archive):
+- """Extract an archive into our tmpdir so that we may inspect it or
+- iterate through its contents for obfuscation
+-
+- Positional arguments:
+-
+- :param archive: An open TarFile object for the archive
+-
+- """
+- if not isinstance(archive, tarfile.TarFile):
+- archive = tarfile.open(archive)
+- path = os.path.join(self.tmpdir, 'cleaner')
+- archive.extractall(path)
+- return os.path.join(path, archive.name.split('/')[-1].split('.tar')[0])
++ for arc in self.archive_types:
++ if arc.check_is_type(self.opts.target):
++ _arc = arc(self.opts.target, self.tmpdir)
++ break
++ if not _arc:
++ return
++ self.report_paths.append(_arc)
++ if _arc.is_nested:
++ self.report_paths.extend(_arc.get_nested_archives())
++ # We need to preserve the top level archive until all
++ # nested archives are processed
++ self.report_paths.remove(_arc)
++ self.nested_archive = _arc
++ if self.nested_archive:
++ self.nested_archive.ui_name = self.nested_archive.description
+
+ def execute(self):
+ """SoSCleaner will begin by inspecting the TARGET option to determine
+@@ -283,6 +271,7 @@ third party.
+ be unpacked, cleaned, and repacked and the final top-level archive will
+ then be repacked as well.
+ """
++ self.arc_name = self.opts.target.split('/')[-1].split('.tar')[0]
+ if self.from_cmdline:
+ self.print_disclaimer()
+ self.report_paths = []
+@@ -290,23 +279,11 @@ third party.
+ self.ui_log.error("Invalid target: no such file or directory %s"
+ % self.opts.target)
+ self._exit(1)
+- if os.path.isdir(self.opts.target):
+- self.arc_name = self.opts.target.split('/')[-1]
+- for _file in os.listdir(self.opts.target):
+- if _file == 'sos_logs':
+- self.report_paths.append(self.opts.target)
+- if (_file.startswith('sosreport') and
+- (_file.endswith(".tar.gz") or _file.endswith(".tar.xz"))):
+- self.report_paths.append(os.path.join(self.opts.target,
+- _file))
+- if not self.report_paths:
+- self.ui_log.error("Invalid target: not an sos directory")
+- self._exit(1)
+- else:
+- self.inspect_target_archive()
++
++ self.inspect_target_archive()
+
+ if not self.report_paths:
+- self.ui_log.error("No valid sos archives or directories found\n")
++ self.ui_log.error("No valid archives or directories found\n")
+ self._exit(1)
+
+ # we have at least one valid target to obfuscate
+@@ -334,33 +311,7 @@ third party.
+
+ final_path = None
+ if len(self.completed_reports) > 1:
+- # we have an archive of archives, so repack the obfuscated tarball
+- arc_name = self.arc_name + '-obfuscated'
+- self.setup_archive(name=arc_name)
+- for arc in self.completed_reports:
+- if arc.is_tarfile:
+- arc_dest = self.obfuscate_string(
+- arc.final_archive_path.split('/')[-1]
+- )
+- self.archive.add_file(arc.final_archive_path,
+- dest=arc_dest)
+- checksum = self.get_new_checksum(arc.final_archive_path)
+- if checksum is not None:
+- dname = self.obfuscate_string(
+- "checksums/%s.%s" % (arc_dest, self.hash_name)
+- )
+- self.archive.add_string(checksum, dest=dname)
+- else:
+- for dirname, dirs, files in os.walk(arc.archive_path):
+- for filename in files:
+- if filename.startswith('sosreport'):
+- continue
+- fname = os.path.join(dirname, filename)
+- dnm = self.obfuscate_string(
+- fname.split(arc.archive_name)[-1].lstrip('/')
+- )
+- self.archive.add_file(fname, dest=dnm)
+- arc_path = self.archive.finalize(self.opts.compression_type)
++ arc_path = self.rebuild_nested_archive()
+ else:
+ arc = self.completed_reports[0]
+ arc_path = arc.final_archive_path
+@@ -371,8 +322,7 @@ third party.
+ )
+ with open(os.path.join(self.sys_tmp, chksum_name), 'w') as cf:
+ cf.write(checksum)
+-
+- self.write_cleaner_log()
++ self.write_cleaner_log()
+
+ final_path = self.obfuscate_string(
+ os.path.join(self.sys_tmp, arc_path.split('/')[-1])
+@@ -393,6 +343,30 @@ third party.
+
+ self.cleanup()
+
++ def rebuild_nested_archive(self):
++ """Handles repacking the nested tarball, now containing only obfuscated
++ copies of the reports, log files, manifest, etc...
++ """
++ # we have an archive of archives, so repack the obfuscated tarball
++ arc_name = self.arc_name + '-obfuscated'
++ self.setup_archive(name=arc_name)
++ for archive in self.completed_reports:
++ arc_dest = archive.final_archive_path.split('/')[-1]
++ checksum = self.get_new_checksum(archive.final_archive_path)
++ if checksum is not None:
++ dname = "checksums/%s.%s" % (arc_dest, self.hash_name)
++ self.archive.add_string(checksum, dest=dname)
++ for dirn, dirs, files in os.walk(self.nested_archive.extracted_path):
++ for filename in files:
++ fname = os.path.join(dirn, filename)
++ dname = fname.split(self.nested_archive.extracted_path)[-1]
++ dname = dname.lstrip('/')
++ self.archive.add_file(fname, dest=dname)
++ # remove it now so we don't balloon our fs space needs
++ os.remove(fname)
++ self.write_cleaner_log(archive=True)
++ return self.archive.finalize(self.opts.compression_type)
++
+ def compile_mapping_dict(self):
+ """Build a dict that contains each parser's map as a key, with the
+ contents as that key's value. This will then be written to disk in the
+@@ -441,7 +415,7 @@ third party.
+ self.log_error("Could not update mapping config file: %s"
+ % err)
+
+- def write_cleaner_log(self):
++ def write_cleaner_log(self, archive=False):
+ """When invoked via the command line, the logging from SoSCleaner will
+ not be added to the archive(s) it processes, so we need to write it
+ separately to disk
+@@ -454,6 +428,10 @@ third party.
+ for line in self.sos_log_file.readlines():
+ logfile.write(line)
+
++ if archive:
++ self.obfuscate_file(log_name)
++ self.archive.add_file(log_name, dest="sos_logs/cleaner.log")
++
+ def get_new_checksum(self, archive_path):
+ """Calculate a new checksum for the obfuscated archive, as the previous
+ checksum will no longer be valid
+@@ -481,11 +459,11 @@ third party.
+ be obfuscated concurrently.
+ """
+ try:
+- if len(self.report_paths) > 1:
+- msg = ("Found %s total reports to obfuscate, processing up to "
+- "%s concurrently\n"
+- % (len(self.report_paths), self.opts.jobs))
+- self.ui_log.info(msg)
++ msg = (
++ "Found %s total reports to obfuscate, processing up to %s "
++ "concurrently\n" % (len(self.report_paths), self.opts.jobs)
++ )
++ self.ui_log.info(msg)
+ if self.opts.keep_binary_files:
+ self.ui_log.warning(
+ "WARNING: binary files that potentially contain sensitive "
+@@ -494,53 +472,67 @@ third party.
+ pool = ThreadPoolExecutor(self.opts.jobs)
+ pool.map(self.obfuscate_report, self.report_paths, chunksize=1)
+ pool.shutdown(wait=True)
++ # finally, obfuscate the nested archive if one exists
++ if self.nested_archive:
++ self._replace_obfuscated_archives()
++ self.obfuscate_report(self.nested_archive)
+ except KeyboardInterrupt:
+ self.ui_log.info("Exiting on user cancel")
+ os._exit(130)
+
++ def _replace_obfuscated_archives(self):
++ """When we have a nested archive, we need to rebuild the original
++ archive, which entails replacing the existing archives with their
++ obfuscated counterparts
++ """
++ for archive in self.completed_reports:
++ os.remove(archive.archive_path)
++ dest = self.nested_archive.extracted_path
++ tarball = archive.final_archive_path.split('/')[-1]
++ dest_name = os.path.join(dest, tarball)
++ shutil.move(archive.final_archive_path, dest)
++ archive.final_archive_path = dest_name
++
+ def preload_all_archives_into_maps(self):
+ """Before doing the actual obfuscation, if we have multiple archives
+ to obfuscate then we need to preload each of them into the mappings
+ to ensure that node1 is obfuscated in node2 as well as node2 being
+ obfuscated in node1's archive.
+ """
+- self.log_info("Pre-loading multiple archives into obfuscation maps")
++ self.log_info("Pre-loading all archives into obfuscation maps")
+ for _arc in self.report_paths:
+- is_dir = os.path.isdir(_arc)
+- if is_dir:
+- _arc_name = _arc
+- else:
+- archive = tarfile.open(_arc)
+- _arc_name = _arc.split('/')[-1].split('.tar')[0]
+- # for each parser, load the map_prep_file into memory, and then
+- # send that for obfuscation. We don't actually obfuscate the file
+- # here, do that in the normal archive loop
+ for _parser in self.parsers:
+- if not _parser.prep_map_file:
++ try:
++ pfile = _arc.prep_files[_parser.name.lower().split()[0]]
++ if not pfile:
++ continue
++ except (IndexError, KeyError):
+ continue
+- if isinstance(_parser.prep_map_file, str):
+- _parser.prep_map_file = [_parser.prep_map_file]
+- for parse_file in _parser.prep_map_file:
+- _arc_path = os.path.join(_arc_name, parse_file)
++ if isinstance(pfile, str):
++ pfile = [pfile]
++ for parse_file in pfile:
++ self.log_debug("Attempting to load %s" % parse_file)
+ try:
+- if is_dir:
+- _pfile = open(_arc_path, 'r')
+- content = _pfile.read()
+- else:
+- _pfile = archive.extractfile(_arc_path)
+- content = _pfile.read().decode('utf-8')
+- _pfile.close()
++ content = _arc.get_file_content(parse_file)
++ if not content:
++ continue
+ if isinstance(_parser, SoSUsernameParser):
+ _parser.load_usernames_into_map(content)
+- for line in content.splitlines():
+- if isinstance(_parser, SoSHostnameParser):
+- _parser.load_hostname_into_map(line)
+- self.obfuscate_line(line)
++ elif isinstance(_parser, SoSHostnameParser):
++ _parser.load_hostname_into_map(
++ content.splitlines()[0]
++ )
++ else:
++ for line in content.splitlines():
++ self.obfuscate_line(line)
+ except Exception as err:
+- self.log_debug("Could not prep %s: %s"
+- % (_arc_path, err))
++ self.log_info(
++ "Could not prepare %s from %s (archive: %s): %s"
++ % (_parser.name, parse_file, _arc.archive_name,
++ err)
++ )
+
+- def obfuscate_report(self, report):
++ def obfuscate_report(self, archive):
+ """Individually handle each archive or directory we've discovered by
+ running through each file therein.
+
+@@ -549,17 +541,12 @@ third party.
+ :param report str: Filepath to the directory or archive
+ """
+ try:
+- if not os.access(report, os.W_OK):
+- msg = "Insufficient permissions on %s" % report
+- self.log_info(msg)
+- self.ui_log.error(msg)
+- return
+-
+- archive = SoSObfuscationArchive(report, self.tmpdir)
+ arc_md = self.cleaner_md.add_section(archive.archive_name)
+ start_time = datetime.now()
+ arc_md.add_field('start_time', start_time)
+- archive.extract()
++ # don't double extract nested archives
++ if not archive.is_extracted:
++ archive.extract()
+ archive.report_msg("Beginning obfuscation...")
+
+ file_list = archive.get_file_list()
+@@ -586,27 +573,28 @@ third party.
+ caller=archive.archive_name)
+
+ # if the archive was already a tarball, repack it
+- method = archive.get_compression()
+- if method:
+- archive.report_msg("Re-compressing...")
+- try:
+- archive.rename_top_dir(
+- self.obfuscate_string(archive.archive_name)
+- )
+- archive.compress(method)
+- except Exception as err:
+- self.log_debug("Archive %s failed to compress: %s"
+- % (archive.archive_name, err))
+- archive.report_msg("Failed to re-compress archive: %s"
+- % err)
+- return
++ if not archive.is_nested:
++ method = archive.get_compression()
++ if method:
++ archive.report_msg("Re-compressing...")
++ try:
++ archive.rename_top_dir(
++ self.obfuscate_string(archive.archive_name)
++ )
++ archive.compress(method)
++ except Exception as err:
++ self.log_debug("Archive %s failed to compress: %s"
++ % (archive.archive_name, err))
++ archive.report_msg("Failed to re-compress archive: %s"
++ % err)
++ return
++ self.completed_reports.append(archive)
+
+ end_time = datetime.now()
+ arc_md.add_field('end_time', end_time)
+ arc_md.add_field('run_time', end_time - start_time)
+ arc_md.add_field('files_obfuscated', len(archive.file_sub_list))
+ arc_md.add_field('total_substitutions', archive.total_sub_count)
+- self.completed_reports.append(archive)
+ rmsg = ''
+ if archive.removed_file_count:
+ rmsg = " [removed %s unprocessable files]"
+@@ -615,7 +603,7 @@ third party.
+
+ except Exception as err:
+ self.ui_log.info("Exception while processing %s: %s"
+- % (report, err))
++ % (archive.archive_name, err))
+
+ def obfuscate_file(self, filename, short_name=None, arc_name=None):
+ """Obfuscate and individual file, line by line.
+@@ -635,6 +623,8 @@ third party.
+ # the requested file doesn't exist in the archive
+ return
+ subs = 0
++ if not short_name:
++ short_name = filename.split('/')[-1]
+ if not os.path.islink(filename):
+ # don't run the obfuscation on the link, but on the actual file
+ # at some other point.
+@@ -745,3 +735,5 @@ third party.
+ for parser in self.parsers:
+ _sec = parse_sec.add_section(parser.name.replace(' ', '_').lower())
+ _sec.add_field('entries', len(parser.mapping.dataset.keys()))
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/cleaner/obfuscation_archive.py b/sos/cleaner/archives/__init__.py
+similarity index 81%
+rename from sos/cleaner/obfuscation_archive.py
+rename to sos/cleaner/archives/__init__.py
+index ea0b7012..795c5a78 100644
+--- a/sos/cleaner/obfuscation_archive.py
++++ b/sos/cleaner/archives/__init__.py
+@@ -40,6 +40,10 @@ class SoSObfuscationArchive():
+ file_sub_list = []
+ total_sub_count = 0
+ removed_file_count = 0
++ type_name = 'undetermined'
++ description = 'undetermined'
++ is_nested = False
++ prep_files = {}
+
+ def __init__(self, archive_path, tmpdir):
+ self.archive_path = archive_path
+@@ -50,7 +54,43 @@ class SoSObfuscationArchive():
+ self.soslog = logging.getLogger('sos')
+ self.ui_log = logging.getLogger('sos_ui')
+ self.skip_list = self._load_skip_list()
+- self.log_info("Loaded %s as an archive" % self.archive_path)
++ self.is_extracted = False
++ self._load_self()
++ self.archive_root = ''
++ self.log_info(
++ "Loaded %s as type %s"
++ % (self.archive_path, self.description)
++ )
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ """Check if the archive is a well-known type we directly support"""
++ return False
++
++ def _load_self(self):
++ if self.is_tarfile:
++ self.tarobj = tarfile.open(self.archive_path)
++
++ def get_nested_archives(self):
++ """Return a list of ObfuscationArchives that represent additional
++ archives found within the target archive. For example, an archive from
++ `sos collect` will return a list of ``SoSReportArchive`` objects.
++
++ This should be overridden by individual types of ObfuscationArchive's
++ """
++ return []
++
++ def get_archive_root(self):
++ """Set the root path for the archive that should be prepended to any
++ filenames given to methods in this class.
++ """
++ if self.is_tarfile:
++ toplevel = self.tarobj.firstmember
++ if toplevel.isdir():
++ return toplevel.name
++ else:
++ return os.sep
++ return os.path.abspath(self.archive_path)
+
+ def report_msg(self, msg):
+ """Helper to easily format ui messages on a per-report basis"""
+@@ -96,10 +136,42 @@ class SoSObfuscationArchive():
+ os.remove(full_fname)
+ self.removed_file_count += 1
+
+- def extract(self):
++ def format_file_name(self, fname):
++ """Based on the type of archive we're dealing with, do whatever that
++ archive requires to a provided **relative** filepath to be able to
++ access it within the archive
++ """
++ if not self.is_extracted:
++ if not self.archive_root:
++ self.archive_root = self.get_archive_root()
++ return os.path.join(self.archive_root, fname)
++ else:
++ return os.path.join(self.extracted_path, fname)
++
++ def get_file_content(self, fname):
++ """Return the content from the specified fname. Particularly useful for
++ tarball-type archives so we can retrieve prep file contents prior to
++ extracting the entire archive
++ """
++ if self.is_extracted is False and self.is_tarfile:
++ filename = self.format_file_name(fname)
++ try:
++ return self.tarobj.extractfile(filename).read().decode('utf-8')
++ except KeyError:
++ self.log_debug(
++ "Unable to retrieve %s: no such file in archive" % fname
++ )
++ return ''
++ else:
++ with open(self.format_file_name(fname), 'r') as to_read:
++ return to_read.read()
++
++ def extract(self, quiet=False):
+ if self.is_tarfile:
+- self.report_msg("Extracting...")
++ if not quiet:
++ self.report_msg("Extracting...")
+ self.extracted_path = self.extract_self()
++ self.is_extracted = True
+ else:
+ self.extracted_path = self.archive_path
+ # if we're running as non-root (e.g. collector), then we can have a
+@@ -317,3 +389,5 @@ class SoSObfuscationArchive():
+ return False
+ except UnicodeDecodeError:
+ return True
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/cleaner/archives/generic.py b/sos/cleaner/archives/generic.py
+new file mode 100644
+index 00000000..2ce6f09b
+--- /dev/null
++++ b/sos/cleaner/archives/generic.py
+@@ -0,0 +1,52 @@
++# Copyright 2020 Red Hat, Inc. Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++
++from sos.cleaner.archives import SoSObfuscationArchive
++
++import os
++import tarfile
++
++
++class DataDirArchive(SoSObfuscationArchive):
++ """A plain directory on the filesystem that is not directly associated with
++ any known or supported collection utility
++ """
++
++ type_name = 'data_dir'
++ description = 'unassociated directory'
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ return os.path.isdir(arc_path)
++
++ def set_archive_root(self):
++ return os.path.abspath(self.archive_path)
++
++
++class TarballArchive(SoSObfuscationArchive):
++ """A generic tar archive that is not associated with any known or supported
++ collection utility
++ """
++
++ type_name = 'tarball'
++ description = 'unassociated tarball'
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ try:
++ return tarfile.is_tarfile(arc_path)
++ except Exception:
++ return False
++
++ def set_archive_root(self):
++ if self.tarobj.firstmember.isdir():
++ return self.tarobj.firstmember.name
++ return ''
+diff --git a/sos/cleaner/archives/sos.py b/sos/cleaner/archives/sos.py
+new file mode 100644
+index 00000000..4401d710
+--- /dev/null
++++ b/sos/cleaner/archives/sos.py
+@@ -0,0 +1,106 @@
++# Copyright 2021 Red Hat, Inc. Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++
++from sos.cleaner.archives import SoSObfuscationArchive
++
++import os
++import tarfile
++
++
++class SoSReportArchive(SoSObfuscationArchive):
++ """This is the class representing an sos report, or in other words the
++ type the archive the SoS project natively generates
++ """
++
++ type_name = 'report'
++ description = 'sos report archive'
++ prep_files = {
++ 'hostname': 'sos_commands/host/hostname',
++ 'ip': 'sos_commands/networking/ip_-o_addr',
++ 'mac': 'sos_commands/networking/ip_-d_address',
++ 'username': [
++ 'sos_commands/login/lastlog_-u_1000-60000',
++ 'sos_commands/login/lastlog_-u_60001-65536',
++ 'sos_commands/login/lastlog_-u_65537-4294967295',
++ # AD users will be reported here, but favor the lastlog files since
++ # those will include local users who have not logged in
++ 'sos_commands/login/last'
++ ]
++ }
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ try:
++ return tarfile.is_tarfile(arc_path) and 'sosreport-' in arc_path
++ except Exception:
++ return False
++
++
++class SoSReportDirectory(SoSReportArchive):
++ """This is the archive class representing a build directory, or in other
++ words what `sos report --clean` will end up using for in-line obfuscation
++ """
++
++ type_name = 'report_dir'
++ description = 'sos report directory'
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ if os.path.isdir(arc_path):
++ return 'sos_logs' in os.listdir(arc_path)
++ return False
++
++
++class SoSCollectorArchive(SoSObfuscationArchive):
++ """Archive class representing the tarball created by ``sos collect``. It
++ will not provide prep files on its own, however it will provide a list
++ of SoSReportArchive's which will then be used to prep the parsers
++ """
++
++ type_name = 'collect'
++ description = 'sos collect tarball'
++ is_nested = True
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ try:
++ return (tarfile.is_tarfile(arc_path) and 'sos-collect' in arc_path)
++ except Exception:
++ return False
++
++ def get_nested_archives(self):
++ self.extract(quiet=True)
++ _path = self.extracted_path
++ archives = []
++ for fname in os.listdir(_path):
++ arc_name = os.path.join(_path, fname)
++ if 'sosreport-' in fname and tarfile.is_tarfile(arc_name):
++ archives.append(SoSReportArchive(arc_name, self.tmpdir))
++ return archives
++
++
++class SoSCollectorDirectory(SoSCollectorArchive):
++ """The archive class representing the temp directory used by ``sos
++ collect`` when ``--clean`` is used during runtime.
++ """
++
++ type_name = 'collect_dir'
++ description = 'sos collect directory'
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ if os.path.isdir(arc_path):
++ for fname in os.listdir(arc_path):
++ if 'sos-collector-' in fname:
++ return True
++ return False
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/cleaner/parsers/__init__.py b/sos/cleaner/parsers/__init__.py
+index af6e375e..e62fd938 100644
+--- a/sos/cleaner/parsers/__init__.py
++++ b/sos/cleaner/parsers/__init__.py
+@@ -37,11 +37,6 @@ class SoSCleanerParser():
+ :cvar map_file_key: The key in the ``map_file`` to read when loading
+ previous obfuscation matches
+ :vartype map_file_key: ``str``
+-
+-
+- :cvar prep_map_file: File to read from an archive to pre-seed the map with
+- matches. E.G. ip_addr for loading IP addresses
+- :vartype prep_map_fie: ``str``
+ """
+
+ name = 'Undefined Parser'
+@@ -49,7 +44,6 @@ class SoSCleanerParser():
+ skip_line_patterns = []
+ skip_files = []
+ map_file_key = 'unset'
+- prep_map_file = []
+
+ def __init__(self, config={}):
+ if self.map_file_key in config:
+diff --git a/sos/cleaner/parsers/hostname_parser.py b/sos/cleaner/parsers/hostname_parser.py
+index 71e13d3f..daa76a62 100644
+--- a/sos/cleaner/parsers/hostname_parser.py
++++ b/sos/cleaner/parsers/hostname_parser.py
+@@ -16,7 +16,6 @@ class SoSHostnameParser(SoSCleanerParser):
+
+ name = 'Hostname Parser'
+ map_file_key = 'hostname_map'
+- prep_map_file = 'sos_commands/host/hostname'
+ regex_patterns = [
+ r'(((\b|_)[a-zA-Z0-9-\.]{1,200}\.[a-zA-Z]{1,63}(\b|_)))'
+ ]
+diff --git a/sos/cleaner/parsers/ip_parser.py b/sos/cleaner/parsers/ip_parser.py
+index 525139e8..71d38be8 100644
+--- a/sos/cleaner/parsers/ip_parser.py
++++ b/sos/cleaner/parsers/ip_parser.py
+@@ -41,7 +41,6 @@ class SoSIPParser(SoSCleanerParser):
+ ]
+
+ map_file_key = 'ip_map'
+- prep_map_file = 'sos_commands/networking/ip_-o_addr'
+
+ def __init__(self, config):
+ self.mapping = SoSIPMap()
+diff --git a/sos/cleaner/parsers/keyword_parser.py b/sos/cleaner/parsers/keyword_parser.py
+index 68de3727..694c6073 100644
+--- a/sos/cleaner/parsers/keyword_parser.py
++++ b/sos/cleaner/parsers/keyword_parser.py
+@@ -20,7 +20,6 @@ class SoSKeywordParser(SoSCleanerParser):
+
+ name = 'Keyword Parser'
+ map_file_key = 'keyword_map'
+- prep_map_file = ''
+
+ def __init__(self, config, keywords=None, keyword_file=None):
+ self.mapping = SoSKeywordMap()
+diff --git a/sos/cleaner/parsers/mac_parser.py b/sos/cleaner/parsers/mac_parser.py
+index 7ca80b8d..c74288cf 100644
+--- a/sos/cleaner/parsers/mac_parser.py
++++ b/sos/cleaner/parsers/mac_parser.py
+@@ -30,7 +30,6 @@ class SoSMacParser(SoSCleanerParser):
+ '534f:53'
+ )
+ map_file_key = 'mac_map'
+- prep_map_file = 'sos_commands/networking/ip_-d_address'
+
+ def __init__(self, config):
+ self.mapping = SoSMacMap()
+diff --git a/sos/cleaner/parsers/username_parser.py b/sos/cleaner/parsers/username_parser.py
+index b142e371..35377a31 100644
+--- a/sos/cleaner/parsers/username_parser.py
++++ b/sos/cleaner/parsers/username_parser.py
+@@ -25,14 +25,6 @@ class SoSUsernameParser(SoSCleanerParser):
+
+ name = 'Username Parser'
+ map_file_key = 'username_map'
+- prep_map_file = [
+- 'sos_commands/login/lastlog_-u_1000-60000',
+- 'sos_commands/login/lastlog_-u_60001-65536',
+- 'sos_commands/login/lastlog_-u_65537-4294967295',
+- # AD users will be reported here, but favor the lastlog files since
+- # those will include local users who have not logged in
+- 'sos_commands/login/last'
+- ]
+ regex_patterns = []
+ skip_list = [
+ 'core',
+diff --git a/tests/cleaner_tests/existing_archive.py b/tests/cleaner_tests/existing_archive.py
+index 0eaf6c8d..e13d1cae 100644
+--- a/tests/cleaner_tests/existing_archive.py
++++ b/tests/cleaner_tests/existing_archive.py
+@@ -28,6 +28,13 @@ class ExistingArchiveCleanTest(StageTwoReportTest):
+ def test_obfuscation_log_created(self):
+ self.assertFileExists(os.path.join(self.tmpdir, '%s-obfuscation.log' % ARCHIVE))
+
++ def test_archive_type_correct(self):
++ with open(os.path.join(self.tmpdir, '%s-obfuscation.log' % ARCHIVE), 'r') as log:
++ for line in log:
++ if "Loaded %s" % ARCHIVE in line:
++ assert 'as type sos report archive' in line, "Incorrect archive type detected: %s" % line
++ break
++
+ def test_from_cmdline_logged(self):
+ with open(os.path.join(self.tmpdir, '%s-obfuscation.log' % ARCHIVE), 'r') as log:
+ for line in log:
+diff --git a/tests/cleaner_tests/full_report_run.py b/tests/cleaner_tests/full_report_run.py
+index 3b28e7a2..2de54946 100644
+--- a/tests/cleaner_tests/full_report_run.py
++++ b/tests/cleaner_tests/full_report_run.py
+@@ -35,6 +35,9 @@ class FullCleanTest(StageTwoReportTest):
+ def test_tarball_named_obfuscated(self):
+ self.assertTrue('obfuscated' in self.archive)
+
++ def test_archive_type_correct(self):
++ self.assertSosLogContains('Loaded .* as type sos report directory')
++
+ def test_hostname_not_in_any_file(self):
+ host = self.sysinfo['pre']['networking']['hostname']
+ # much faster to just use grep here
+diff --git a/tests/cleaner_tests/report_with_mask.py b/tests/cleaner_tests/report_with_mask.py
+index 4f94ba33..08e873d4 100644
+--- a/tests/cleaner_tests/report_with_mask.py
++++ b/tests/cleaner_tests/report_with_mask.py
+@@ -31,6 +31,9 @@ class ReportWithMask(StageOneReportTest):
+ def test_tarball_named_obfuscated(self):
+ self.assertTrue('obfuscated' in self.archive)
+
++ def test_archive_type_correct(self):
++ self.assertSosLogContains('Loaded .* as type sos report directory')
++
+ def test_localhost_was_obfuscated(self):
+ self.assertFileHasContent('/etc/hostname', 'host0')
+
+--
+2.31.1
+
+From 9b119f860eaec089f7ef884ff39c42589a662994 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 1 Sep 2021 00:34:04 -0400
+Subject: [PATCH] [hostname_map] Add a catch for single-character hostnames
+
+If a log file was truncated at a specific boundary in a string of the
+FQDN of the host such that we only get a couple characters before the
+rest of the domain, we would previously bodly replace all instances of
+that character with the obfuscated short name; not very helpful.
+
+Instead, don't sanitize the short name if this happens and instead
+obfuscate the whole FQDN as 'unknown.example.com'.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/hostname_map.py | 9 ++++++++-
+ 1 file changed, 8 insertions(+), 1 deletion(-)
+
+diff --git a/sos/cleaner/mappings/hostname_map.py b/sos/cleaner/mappings/hostname_map.py
+index d4b2c88e..e70a5530 100644
+--- a/sos/cleaner/mappings/hostname_map.py
++++ b/sos/cleaner/mappings/hostname_map.py
+@@ -184,7 +184,14 @@ class SoSHostnameMap(SoSMap):
+ hostname = host[0]
+ domain = host[1:]
+ # obfuscate the short name
+- ob_hostname = self.sanitize_short_name(hostname)
++ if len(hostname) > 2:
++ ob_hostname = self.sanitize_short_name(hostname)
++ else:
++ # by best practice it appears the host part of the fqdn was cut
++ # off due to some form of truncating, as such don't obfuscate
++ # short strings that are likely to throw off obfuscation of
++ # unrelated bits and paths
++ ob_hostname = 'unknown'
+ ob_domain = self.sanitize_domain(domain)
+ self.dataset[item] = ob_domain
+ return '.'.join([ob_hostname, ob_domain])
+--
+2.31.1
+
+From f3f3e763d7c31b7b7cafdf8dd4dab87056fb7696 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 1 Sep 2021 15:54:55 -0400
+Subject: [PATCH] [cleaner] Add support for Insights client archives
+
+Adds a new type of `SoSObfuscationArchive` to add support for
+obfuscating archives generated by the Insights project.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ man/en/sos-clean.1 | 1 +
+ sos/cleaner/__init__.py | 4 ++-
+ sos/cleaner/archives/insights.py | 42 ++++++++++++++++++++++++++++++++
+ 3 files changed, 46 insertions(+), 1 deletion(-)
+ create mode 100644 sos/cleaner/archives/insights.py
+
+diff --git a/man/en/sos-clean.1 b/man/en/sos-clean.1
+index 54026713..358ec0cb 100644
+--- a/man/en/sos-clean.1
++++ b/man/en/sos-clean.1
+@@ -105,6 +105,7 @@ The following are accepted values for this option:
+ \fBauto\fR Automatically detect the archive type
+ \fBreport\fR An archive generated by \fBsos report\fR
+ \fBcollect\fR An archive generated by \fBsos collect\fR
++ \fBinsights\fR An archive generated by the \fBinsights-client\fR package
+
+ The following may also be used, however note that these do not attempt to pre-load
+ any information from the archives into the parsers. This means that, among other limitations,
+diff --git a/sos/cleaner/__init__.py b/sos/cleaner/__init__.py
+index 6d2eb483..3e08aa28 100644
+--- a/sos/cleaner/__init__.py
++++ b/sos/cleaner/__init__.py
+@@ -29,6 +29,7 @@ from sos.cleaner.archives.sos import (SoSReportArchive, SoSReportDirectory,
+ SoSCollectorArchive,
+ SoSCollectorDirectory)
+ from sos.cleaner.archives.generic import DataDirArchive, TarballArchive
++from sos.cleaner.archives.insights import InsightsArchive
+ from sos.utilities import get_human_readable
+ from textwrap import fill
+
+@@ -100,6 +101,7 @@ class SoSCleaner(SoSComponent):
+ SoSReportArchive,
+ SoSCollectorDirectory,
+ SoSCollectorArchive,
++ InsightsArchive,
+ # make sure these two are always last as they are fallbacks
+ DataDirArchive,
+ TarballArchive
+@@ -194,7 +196,7 @@ third party.
+ help='The directory or archive to obfuscate')
+ clean_grp.add_argument('--archive-type', default='auto',
+ choices=['auto', 'report', 'collect',
+- 'data-dir', 'tarball'],
++ 'insights', 'data-dir', 'tarball'],
+ help=('Specify what kind of archive the target '
+ 'was generated as'))
+ clean_grp.add_argument('--domains', action='extend', default=[],
+diff --git a/sos/cleaner/archives/insights.py b/sos/cleaner/archives/insights.py
+new file mode 100644
+index 00000000..dab48b16
+--- /dev/null
++++ b/sos/cleaner/archives/insights.py
+@@ -0,0 +1,42 @@
++# Copyright 2021 Red Hat, Inc. Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++
++from sos.cleaner.archives import SoSObfuscationArchive
++
++import tarfile
++
++
++class InsightsArchive(SoSObfuscationArchive):
++ """This class represents archives generated by the insights-client utility
++ for RHEL systems.
++ """
++
++ type_name = 'insights'
++ description = 'insights-client archive'
++
++ prep_files = {
++ 'hostname': 'data/insights_commands/hostname_-f',
++ 'ip': 'data/insights_commands/ip_addr',
++ 'mac': 'data/insights_commands/ip_addr'
++ }
++
++ @classmethod
++ def check_is_type(cls, arc_path):
++ try:
++ return tarfile.is_tarfile(arc_path) and 'insights-' in arc_path
++ except Exception:
++ return False
++
++ def get_archive_root(self):
++ top = self.archive_path.split('/')[-1].split('.tar')[0]
++ if self.tarobj.firstmember.name == '.':
++ top = './' + top
++ return top
+--
+2.31.1
+
+From 9639dc3d240076b55f2a1d04b43ea42bebd09215 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 16 Nov 2021 17:50:42 -0500
+Subject: [PATCH] [clean,hostname_parser] Source /etc/hosts for obfuscation
+
+Up until now, our sourcing of hostnames/domains for obfuscation has been
+dependent upon the output of the `hostname` command. However, some
+scenarios have come up where sourcing `/etc/hosts` is advantageous for
+several reasons:
+
+First, if `hostname` output is unavailable, this provides a fallback
+measure.
+
+Second, `/etc/hosts` is a common place to have short names defined which
+would otherwise not be detected (or at the very least would result in a
+race condition based on where/if the short name was elsewhere able to be
+gleaned from an FQDN), thus leaving the potential for unobfuscated data
+in an archive.
+
+Due to both the nature of hostname obfuscation and the malleable syntax
+of `/etc/hosts`, the parsing of this file needs special handling not
+covered by our more generic parsing and obfuscation methods.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/__init__.py | 11 ++++++++---
+ sos/cleaner/archives/sos.py | 5 ++++-
+ sos/cleaner/parsers/hostname_parser.py | 19 +++++++++++++++++++
+ 3 files changed, 31 insertions(+), 4 deletions(-)
+
+diff --git a/sos/cleaner/__init__.py b/sos/cleaner/__init__.py
+index ed461a8f..3f530d44 100644
+--- a/sos/cleaner/__init__.py
++++ b/sos/cleaner/__init__.py
+@@ -523,9 +523,14 @@ third party.
+ if isinstance(_parser, SoSUsernameParser):
+ _parser.load_usernames_into_map(content)
+ elif isinstance(_parser, SoSHostnameParser):
+- _parser.load_hostname_into_map(
+- content.splitlines()[0]
+- )
++ if 'hostname' in parse_file:
++ _parser.load_hostname_into_map(
++ content.splitlines()[0]
++ )
++ elif 'etc/hosts' in parse_file:
++ _parser.load_hostname_from_etc_hosts(
++ content
++ )
+ else:
+ for line in content.splitlines():
+ self.obfuscate_line(line)
+diff --git a/sos/cleaner/archives/sos.py b/sos/cleaner/archives/sos.py
+index 4401d710..f8720c88 100644
+--- a/sos/cleaner/archives/sos.py
++++ b/sos/cleaner/archives/sos.py
+@@ -23,7 +23,10 @@ class SoSReportArchive(SoSObfuscationArchive):
+ type_name = 'report'
+ description = 'sos report archive'
+ prep_files = {
+- 'hostname': 'sos_commands/host/hostname',
++ 'hostname': [
++ 'sos_commands/host/hostname',
++ 'etc/hosts'
++ ],
+ 'ip': 'sos_commands/networking/ip_-o_addr',
+ 'mac': 'sos_commands/networking/ip_-d_address',
+ 'username': [
+diff --git a/sos/cleaner/parsers/hostname_parser.py b/sos/cleaner/parsers/hostname_parser.py
+index daa76a62..0a733bee 100644
+--- a/sos/cleaner/parsers/hostname_parser.py
++++ b/sos/cleaner/parsers/hostname_parser.py
+@@ -61,6 +61,25 @@ class SoSHostnameParser(SoSCleanerParser):
+ self.mapping.add(high_domain)
+ self.mapping.add(hostname_string)
+
++ def load_hostname_from_etc_hosts(self, content):
++ """Parse an archive's copy of /etc/hosts, which requires handling that
++ is separate from the output of the `hostname` command. Just like
++ load_hostname_into_map(), this has to be done explicitly and we
++ cannot rely upon the more generic methods to do this reliably.
++ """
++ lines = content.splitlines()
++ for line in lines:
++ if line.startswith('#') or 'localhost' in line:
++ continue
++ hostln = line.split()[1:]
++ for host in hostln:
++ if len(host.split('.')) == 1:
++ # only generate a mapping for fqdns but still record the
++ # short name here for later obfuscation with parse_line()
++ self.short_names.append(host)
++ else:
++ self.mapping.add(host)
++
+ def parse_line(self, line):
+ """Override the default parse_line() method to also check for the
+ shortname of the host derived from the hostname.
+--
+2.31.1
+
+From c1680226b53452b18f27f2e76c3e0e03e521f935 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 17 Nov 2021 13:11:33 -0500
+Subject: [PATCH] [clean, hostname] Fix unintentionally case sensitive
+ shortname handling
+
+It was discovered that our extra handling for shortnames was
+unintentionally case sensitive. Fix this to ensure that shortnames are
+obfuscated regardless of case in all collected text.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/hostname_map.py | 6 +++---
+ sos/cleaner/parsers/hostname_parser.py | 8 +++++---
+ tests/cleaner_tests/full_report_run.py | 21 ++++++++++++++++++++-
+ 3 files changed, 28 insertions(+), 7 deletions(-)
+
+diff --git a/sos/cleaner/mappings/hostname_map.py b/sos/cleaner/mappings/hostname_map.py
+index e70a5530..0fe78fb1 100644
+--- a/sos/cleaner/mappings/hostname_map.py
++++ b/sos/cleaner/mappings/hostname_map.py
+@@ -169,13 +169,13 @@ class SoSHostnameMap(SoSMap):
+
+ def sanitize_item(self, item):
+ host = item.split('.')
+- if all([h.isupper() for h in host]):
++ if len(host) > 1 and all([h.isupper() for h in host]):
+ # by convention we have just a domain
+ _host = [h.lower() for h in host]
+ return self.sanitize_domain(_host).upper()
+ if len(host) == 1:
+ # we have a shortname for a host
+- return self.sanitize_short_name(host[0])
++ return self.sanitize_short_name(host[0].lower())
+ if len(host) == 2:
+ # we have just a domain name, e.g. example.com
+ return self.sanitize_domain(host)
+@@ -185,7 +185,7 @@ class SoSHostnameMap(SoSMap):
+ domain = host[1:]
+ # obfuscate the short name
+ if len(hostname) > 2:
+- ob_hostname = self.sanitize_short_name(hostname)
++ ob_hostname = self.sanitize_short_name(hostname.lower())
+ else:
+ # by best practice it appears the host part of the fqdn was cut
+ # off due to some form of truncating, as such don't obfuscate
+diff --git a/sos/cleaner/parsers/hostname_parser.py b/sos/cleaner/parsers/hostname_parser.py
+index 0a733bee..7fd0e698 100644
+--- a/sos/cleaner/parsers/hostname_parser.py
++++ b/sos/cleaner/parsers/hostname_parser.py
+@@ -8,6 +8,8 @@
+ #
+ # See the LICENSE file in the source distribution for further information.
+
++import re
++
+ from sos.cleaner.parsers import SoSCleanerParser
+ from sos.cleaner.mappings.hostname_map import SoSHostnameMap
+
+@@ -91,9 +93,9 @@ class SoSHostnameParser(SoSCleanerParser):
+ """
+ if search in self.mapping.skip_keys:
+ return ln, count
+- if search in ln:
+- count += ln.count(search)
+- ln = ln.replace(search, self.mapping.get(repl or search))
++ _reg = re.compile(search, re.I)
++ if _reg.search(ln):
++ return _reg.subn(self.mapping.get(repl or search), ln)
+ return ln, count
+
+ count = 0
+diff --git a/tests/cleaner_tests/full_report_run.py b/tests/cleaner_tests/full_report_run.py
+index 2de54946..0b23acaf 100644
+--- a/tests/cleaner_tests/full_report_run.py
++++ b/tests/cleaner_tests/full_report_run.py
+@@ -26,6 +26,24 @@ class FullCleanTest(StageTwoReportTest):
+ # replace with an empty placeholder, make sure that this test case is not
+ # influenced by previous clean runs
+ files = ['/etc/sos/cleaner/default_mapping']
++ packages = {
++ 'rhel': ['python3-systemd'],
++ 'ubuntu': ['python3-systemd']
++ }
++
++ def pre_sos_setup(self):
++ # ensure that case-insensitive matching of FQDNs and shortnames work
++ from systemd import journal
++ from socket import gethostname
++ host = gethostname()
++ short = host.split('.')[0]
++ sosfd = journal.stream('sos-testing')
++ sosfd.write(
++ "This is a test line from sos clean testing. The hostname %s "
++ "should not appear, nor should %s in an obfuscated archive. The "
++ "shortnames of %s and %s should also not appear."
++ % (host.lower(), host.upper(), short.lower(), short.upper())
++ )
+
+ def test_private_map_was_generated(self):
+ self.assertOutputContains('A mapping of obfuscated elements is available at')
+@@ -40,8 +58,9 @@ class FullCleanTest(StageTwoReportTest):
+
+ def test_hostname_not_in_any_file(self):
+ host = self.sysinfo['pre']['networking']['hostname']
++ short = host.split('.')[0]
+ # much faster to just use grep here
+- content = self.grep_for_content(host)
++ content = self.grep_for_content(host) + self.grep_for_content(short)
+ if not content:
+ assert True
+ else:
+--
+2.31.1
+
+From aaeb8cb57ed55598ab744b96d4f127aedebcb292 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 21 Sep 2021 15:23:20 -0400
+Subject: [PATCH] [build] Add archives to setup.py packages
+
+Adds the newly abstracted `sos.cleaner.archives` package to `setup.py`
+so that manual builds will properly include it.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ setup.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/setup.py b/setup.py
+index 1e8d8e2dc5..7653b59de3 100644
+--- a/setup.py
++++ b/setup.py
+@@ -102,7 +102,7 @@ def copy_file (self, filename, dirname):
+ 'sos.policies.package_managers', 'sos.policies.init_systems',
+ 'sos.report', 'sos.report.plugins', 'sos.collector',
+ 'sos.collector.clusters', 'sos.cleaner', 'sos.cleaner.mappings',
+- 'sos.cleaner.parsers'
++ 'sos.cleaner.parsers', 'sos.cleaner.archives'
+ ],
+ cmdclass=cmdclass,
+ command_options=command_options,
+--
+2.31.1
+
+From ba3528230256429a4394f155a9ca1fdb91cf3560 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 30 Nov 2021 12:46:34 -0500
+Subject: [PATCH 1/2] [hostname] Simplify case matching for domains
+
+Instead of special handling all uppercase domain conventions, use our
+normal flow for obfuscation and just match the casing at the end of the
+sanitization routine.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/hostname_map.py | 14 ++++++++------
+ 1 file changed, 8 insertions(+), 6 deletions(-)
+
+diff --git a/sos/cleaner/mappings/hostname_map.py b/sos/cleaner/mappings/hostname_map.py
+index 0fe78fb1..5cd8e985 100644
+--- a/sos/cleaner/mappings/hostname_map.py
++++ b/sos/cleaner/mappings/hostname_map.py
+@@ -169,16 +169,15 @@ class SoSHostnameMap(SoSMap):
+
+ def sanitize_item(self, item):
+ host = item.split('.')
+- if len(host) > 1 and all([h.isupper() for h in host]):
+- # by convention we have just a domain
+- _host = [h.lower() for h in host]
+- return self.sanitize_domain(_host).upper()
+ if len(host) == 1:
+ # we have a shortname for a host
+ return self.sanitize_short_name(host[0].lower())
+ if len(host) == 2:
+ # we have just a domain name, e.g. example.com
+- return self.sanitize_domain(host)
++ dname = self.sanitize_domain(host)
++ if all([h.isupper() for h in host]):
++ dname = dname.upper()
++ return dname
+ if len(host) > 2:
+ # we have an FQDN, e.g. foo.example.com
+ hostname = host[0]
+@@ -194,7 +193,10 @@ class SoSHostnameMap(SoSMap):
+ ob_hostname = 'unknown'
+ ob_domain = self.sanitize_domain(domain)
+ self.dataset[item] = ob_domain
+- return '.'.join([ob_hostname, ob_domain])
++ _fqdn = '.'.join([ob_hostname, ob_domain])
++ if all([h.isupper() for h in host]):
++ _fqdn = _fqdn.upper()
++ return _fqdn
+
+ def sanitize_short_name(self, hostname):
+ """Obfuscate the short name of the host with an incremented counter
+--
+2.31.1
+
+
+From 189586728de22dd55122c1f7e06b19590f9a788f Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 30 Nov 2021 12:47:58 -0500
+Subject: [PATCH 2/2] [username] Improve username sourcing and remove case
+ sensitivity
+
+First, don't skip the first line of `last` output, and instead add the
+header from lastlog to the skip list. Additionally, add
+`/etc/cron.allow` and `/etc/cron.deny` as sources for usernames that
+might not appear in other locations in certain environments.
+
+Also, make matching and replacement case insensitive.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/archives/sos.py | 4 +++-
+ sos/cleaner/mappings/username_map.py | 2 +-
+ sos/cleaner/parsers/username_parser.py | 14 +++++++++-----
+ 3 files changed, 13 insertions(+), 7 deletions(-)
+
+diff --git a/sos/cleaner/archives/sos.py b/sos/cleaner/archives/sos.py
+index f8720c88..12766496 100644
+--- a/sos/cleaner/archives/sos.py
++++ b/sos/cleaner/archives/sos.py
+@@ -35,7 +35,9 @@ class SoSReportArchive(SoSObfuscationArchive):
+ 'sos_commands/login/lastlog_-u_65537-4294967295',
+ # AD users will be reported here, but favor the lastlog files since
+ # those will include local users who have not logged in
+- 'sos_commands/login/last'
++ 'sos_commands/login/last',
++ 'etc/cron.allow',
++ 'etc/cron.deny'
+ ]
+ }
+
+diff --git a/sos/cleaner/mappings/username_map.py b/sos/cleaner/mappings/username_map.py
+index cdbf36fe..7ecccd7b 100644
+--- a/sos/cleaner/mappings/username_map.py
++++ b/sos/cleaner/mappings/username_map.py
+@@ -33,5 +33,5 @@ class SoSUsernameMap(SoSMap):
+ ob_name = "obfuscateduser%s" % self.name_count
+ self.name_count += 1
+ if ob_name in self.dataset.values():
+- return self.sanitize_item(username)
++ return self.sanitize_item(username.lower())
+ return ob_name
+diff --git a/sos/cleaner/parsers/username_parser.py b/sos/cleaner/parsers/username_parser.py
+index 35377a31..229c7de4 100644
+--- a/sos/cleaner/parsers/username_parser.py
++++ b/sos/cleaner/parsers/username_parser.py
+@@ -8,6 +8,7 @@
+ #
+ # See the LICENSE file in the source distribution for further information.
+
++import re
+
+ from sos.cleaner.parsers import SoSCleanerParser
+ from sos.cleaner.mappings.username_map import SoSUsernameMap
+@@ -34,6 +35,7 @@ class SoSUsernameParser(SoSCleanerParser):
+ 'reboot',
+ 'root',
+ 'ubuntu',
++ 'username',
+ 'wtmp'
+ ]
+
+@@ -47,12 +49,12 @@ class SoSUsernameParser(SoSCleanerParser):
+ this parser, we need to override the initial parser prepping here.
+ """
+ users = set()
+- for line in content.splitlines()[1:]:
++ for line in content.splitlines():
+ try:
+ user = line.split()[0]
+ except Exception:
+ continue
+- if user in self.skip_list:
++ if user.lower() in self.skip_list:
+ continue
+ users.add(user)
+ for each in users:
+@@ -61,7 +63,9 @@ class SoSUsernameParser(SoSCleanerParser):
+ def parse_line(self, line):
+ count = 0
+ for username in sorted(self.mapping.dataset.keys(), reverse=True):
+- if username in line:
+- count = line.count(username)
+- line = line.replace(username, self.mapping.get(username))
++ _reg = re.compile(username, re.I)
++ if _reg.search(line):
++ line, count = _reg.subn(
++ self.mapping.get(username.lower()), line
++ )
+ return line, count
+--
+2.31.1
+
+From cafd0f3a52436a3966576e7db21e5dd17c06f0cc Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Sun, 12 Dec 2021 11:10:46 -0500
+Subject: [PATCH] [hostname] Fix edge case for new hosts in a known subdomain
+
+Fixes an edge case that would cause us to at first not recognize that a
+given hostname string is a new host in a known subdomain, but then on
+the obfuscation attempt properly recognize it as such and result in an
+incomplete obfuscation.
+
+This was mostly triggered by specific patterns for build hosts within
+`sos_commands/rpm/package-data`. With this refined check, these types of
+matches are properly obfuscated.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/hostname_map.py | 9 +++++----
+ 1 file changed, 5 insertions(+), 4 deletions(-)
+
+diff --git a/sos/cleaner/mappings/hostname_map.py b/sos/cleaner/mappings/hostname_map.py
+index 5cd8e9857..33b0e6c80 100644
+--- a/sos/cleaner/mappings/hostname_map.py
++++ b/sos/cleaner/mappings/hostname_map.py
+@@ -129,7 +129,7 @@ def get(self, item):
+ item = item[0:-1]
+ if not self.domain_name_in_loaded_domains(item.lower()):
+ return item
+- if item.endswith(('.yaml', '.yml', '.crt', '.key', '.pem')):
++ if item.endswith(('.yaml', '.yml', '.crt', '.key', '.pem', '.log')):
+ ext = '.' + item.split('.')[-1]
+ item = item.replace(ext, '')
+ suffix += ext
+@@ -148,7 +148,8 @@ def get(self, item):
+ if len(_test) == 1 or not _test[0]:
+ # does not match existing obfuscation
+ continue
+- elif _test[0].endswith('.') and not _host_substr:
++ elif not _host_substr and (_test[0].endswith('.') or
++ item.endswith(_existing)):
+ # new hostname in known domain
+ final = super(SoSHostnameMap, self).get(item)
+ break
+@@ -219,8 +220,8 @@ def sanitize_domain(self, domain):
+ # don't obfuscate vendor domains
+ if re.match(_skip, '.'.join(domain)):
+ return '.'.join(domain)
+- top_domain = domain[-1]
+- dname = '.'.join(domain[0:-1])
++ top_domain = domain[-1].lower()
++ dname = '.'.join(domain[0:-1]).lower()
+ ob_domain = self._new_obfuscated_domain(dname)
+ ob_domain = '.'.join([ob_domain, top_domain])
+ self.dataset['.'.join(domain)] = ob_domain
+From f5e1298162a9393ea2d9f5c4df40dfece50f5f88 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 6 Jan 2022 13:15:15 -0500
+Subject: [PATCH 1/3] [hostname] Fix loading and detection of long base domains
+
+Our domain matching has up to now assumed that users would be providing
+'base' domains such as 'example.com' whereby something like
+'foo.bar.example.com' is a subdomain (or host) within that base domain.
+
+However, the use case exists to provide 'foo.bar.example.com' as the
+base domain, without wanting to obfuscate 'example.com' directly.
+
+This commit fixes our handling of both loading these longer domains and
+doing the 'domain is part of a domain we want to obfuscate' check.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/hostname_map.py | 9 ++++++++-
+ 1 file changed, 8 insertions(+), 1 deletion(-)
+
+diff --git a/sos/cleaner/mappings/hostname_map.py b/sos/cleaner/mappings/hostname_map.py
+index 33b0e6c8..7a7cf6b8 100644
+--- a/sos/cleaner/mappings/hostname_map.py
++++ b/sos/cleaner/mappings/hostname_map.py
+@@ -50,10 +50,14 @@ class SoSHostnameMap(SoSMap):
+ in this parser, we need to re-inject entries from the map_file into
+ these dicts and not just the underlying 'dataset' dict
+ """
+- for domain in self.dataset:
++ for domain, ob_pair in self.dataset.items():
+ if len(domain.split('.')) == 1:
+ self.hosts[domain.split('.')[0]] = self.dataset[domain]
+ else:
++ if ob_pair.startswith('obfuscateddomain'):
++ # directly exact domain matches
++ self._domains[domain] = ob_pair.split('.')[0]
++ continue
+ # strip the host name and trailing top-level domain so that
+ # we in inject the domain properly for later string matching
+
+@@ -102,9 +106,12 @@ class SoSHostnameMap(SoSMap):
+ and should be obfuscated
+ """
+ host = domain.split('.')
++ no_tld = '.'.join(domain.split('.')[0:-1])
+ if len(host) == 1:
+ # don't block on host's shortname
+ return host[0] in self.hosts.keys()
++ elif any([no_tld.endswith(_d) for _d in self._domains]):
++ return True
+ else:
+ domain = host[0:-1]
+ for known_domain in self._domains:
+--
+2.31.1
+
+
+From e241cf33a14ecd4e848a5fd857c5d3d7d07fbd71 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 6 Jan 2022 13:18:44 -0500
+Subject: [PATCH 2/3] [cleaner] Improve parser-specific file skipping
+
+This commit improves our handling of skipping files on a per-parser
+basis, by first filtering the list of parsers that `obfuscate_line()`
+will iterate over by the parser's `skip_file` class attr, rather than
+relying on higher-level checks.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/__init__.py | 17 ++++++++++++++---
+ 1 file changed, 14 insertions(+), 3 deletions(-)
+
+diff --git a/sos/cleaner/__init__.py b/sos/cleaner/__init__.py
+index 3f530d44..5686e213 100644
+--- a/sos/cleaner/__init__.py
++++ b/sos/cleaner/__init__.py
+@@ -12,6 +12,7 @@ import hashlib
+ import json
+ import logging
+ import os
++import re
+ import shutil
+ import tempfile
+
+@@ -640,10 +641,16 @@ third party.
+ self.log_debug("Obfuscating %s" % short_name or filename,
+ caller=arc_name)
+ tfile = tempfile.NamedTemporaryFile(mode='w', dir=self.tmpdir)
++ _parsers = [
++ _p for _p in self.parsers if not
++ any([
++ re.match(p, short_name) for p in _p.skip_files
++ ])
++ ]
+ with open(filename, 'r') as fname:
+ for line in fname:
+ try:
+- line, count = self.obfuscate_line(line)
++ line, count = self.obfuscate_line(line, _parsers)
+ subs += count
+ tfile.write(line)
+ except Exception as err:
+@@ -713,7 +720,7 @@ third party.
+ pass
+ return string_data
+
+- def obfuscate_line(self, line):
++ def obfuscate_line(self, line, parsers=None):
+ """Run a line through each of the obfuscation parsers, keeping a
+ cumulative total of substitutions done on that particular line.
+
+@@ -721,6 +728,8 @@ third party.
+
+ :param line str: The raw line as read from the file being
+ processed
++ :param parsers: A list of parser objects to obfuscate
++ with. If None, use all.
+
+ Returns the fully obfuscated line and the number of substitutions made
+ """
+@@ -729,7 +738,9 @@ third party.
+ count = 0
+ if not line.strip():
+ return line, count
+- for parser in self.parsers:
++ if parsers is None:
++ parsers = self.parsers
++ for parser in parsers:
+ try:
+ line, _count = parser.parse_line(line)
+ count += _count
+--
+2.31.1
+
+
+From 96c9a833e77639a853b7d3d6f1df68bbbbe5e9cb Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 6 Jan 2022 13:20:32 -0500
+Subject: [PATCH 3/3] [cleaner] Add skips for known files and usernames
+
+Adds skips for `/proc/kallsyms` which should never be obfuscated, as
+well as any packaging-related log file for the IP parser. Further, do
+not obfuscate the `stack` users, as that is a well-known user for many
+configurations that, if obfuscated, could result in undesired string
+substitutions in normal logging.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/archives/__init__.py | 2 ++
+ sos/cleaner/parsers/ip_parser.py | 3 ++-
+ sos/cleaner/parsers/username_parser.py | 1 +
+ 3 files changed, 5 insertions(+), 1 deletion(-)
+
+diff --git a/sos/cleaner/archives/__init__.py b/sos/cleaner/archives/__init__.py
+index 795c5a78..cbf1f809 100644
+--- a/sos/cleaner/archives/__init__.py
++++ b/sos/cleaner/archives/__init__.py
+@@ -43,6 +43,7 @@ class SoSObfuscationArchive():
+ type_name = 'undetermined'
+ description = 'undetermined'
+ is_nested = False
++ skip_files = []
+ prep_files = {}
+
+ def __init__(self, archive_path, tmpdir):
+@@ -111,6 +112,7 @@ class SoSObfuscationArchive():
+ Returns: list of files and file regexes
+ """
+ return [
++ 'proc/kallsyms',
+ 'sosreport-',
+ 'sys/firmware',
+ 'sys/fs',
+diff --git a/sos/cleaner/parsers/ip_parser.py b/sos/cleaner/parsers/ip_parser.py
+index 71d38be8..b007368c 100644
+--- a/sos/cleaner/parsers/ip_parser.py
++++ b/sos/cleaner/parsers/ip_parser.py
+@@ -37,7 +37,8 @@ class SoSIPParser(SoSCleanerParser):
+ 'sos_commands/snappy/snap_list_--all',
+ 'sos_commands/snappy/snap_--version',
+ 'sos_commands/vulkan/vulkaninfo',
+- 'var/log/.*dnf.*'
++ 'var/log/.*dnf.*',
++ 'var/log/.*packag.*' # get 'packages' and 'packaging' logs
+ ]
+
+ map_file_key = 'ip_map'
+diff --git a/sos/cleaner/parsers/username_parser.py b/sos/cleaner/parsers/username_parser.py
+index 229c7de4..3208a655 100644
+--- a/sos/cleaner/parsers/username_parser.py
++++ b/sos/cleaner/parsers/username_parser.py
+@@ -32,6 +32,7 @@ class SoSUsernameParser(SoSCleanerParser):
+ 'nobody',
+ 'nfsnobody',
+ 'shutdown',
++ 'stack',
+ 'reboot',
+ 'root',
+ 'ubuntu',
+--
+2.31.1
+
+From 7ebb2ce0bcd13c1b3aada648aceb20b5aff636d9 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 15 Feb 2022 14:18:02 -0500
+Subject: [PATCH] [host] Skip entire /etc/sos/cleaner directory
+
+While `default_mapping` is typically the only file expected under
+`/etc/sos/cleaner/` it is possible for other mapping files (such as
+backups) to appear there.
+
+Make the `add_forbidden_path()` spec here target the entire cleaner
+directory to avoid ever capturing these map files.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/host.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/host.py b/sos/report/plugins/host.py
+index 5e21da7b8e..95a3b9cd95 100644
+--- a/sos/report/plugins/host.py
++++ b/sos/report/plugins/host.py
+@@ -20,7 +20,7 @@ class Host(Plugin, IndependentPlugin):
+
+ def setup(self):
+
+- self.add_forbidden_path('/etc/sos/cleaner/default_mapping')
++ self.add_forbidden_path('/etc/sos/cleaner')
+
+ self.add_cmd_output('hostname', root_symlink='hostname')
+ self.add_cmd_output('uptime', root_symlink='uptime')
diff --git a/SOURCES/sos-bz2025611-RHTS-api-change.patch b/SOURCES/sos-bz2025611-RHTS-api-change.patch
new file mode 100644
index 0000000..580117f
--- /dev/null
+++ b/SOURCES/sos-bz2025611-RHTS-api-change.patch
@@ -0,0 +1,224 @@
+From 2e8b5e2d4f30854cce93d149fc7d24b9d9cfd02c Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 19 Nov 2021 16:16:07 +0100
+Subject: [PATCH 1/3] [policies] strip path from SFTP upload filename
+
+When case_id is not supplied, we ask SFTP server to store the uploaded
+file under name /var/tmp/<tarball>, which is confusing.
+
+Let remove the path from it also in case_id not supplied.
+
+Related to: #2764
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/redhat.py | 6 +++---
+ 1 file changed, 3 insertions(+), 3 deletions(-)
+
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index 3476e21fb..8817fc785 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -269,10 +269,10 @@ def _get_sftp_upload_name(self):
+ """The RH SFTP server will only automatically connect file uploads to
+ cases if the filename _starts_ with the case number
+ """
++ fname = self.upload_archive_name.split('/')[-1]
+ if self.case_id:
+- return "%s_%s" % (self.case_id,
+- self.upload_archive_name.split('/')[-1])
+- return self.upload_archive_name
++ return "%s_%s" % (self.case_id, fname)
++ return fname
+
+ def upload_sftp(self):
+ """Override the base upload_sftp to allow for setting an on-demand
+
+From 61023b29a656dd7afaa4a0643368b0a53f1a3779 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 19 Nov 2021 17:31:31 +0100
+Subject: [PATCH 2/3] [redhat] update SFTP API version to v2
+
+Change API version from v1 to v2, which includes:
+- change of URL
+- different URI
+- POST method for token generation instead of GET
+
+Resolves: #2764
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/redhat.py | 10 +++++-----
+ 1 file changed, 5 insertions(+), 5 deletions(-)
+
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index 8817fc785..e4e2b8835 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -175,7 +175,7 @@ def get_tmp_dir(self, opt_tmp_dir):
+ No changes will be made to system configuration.
+ """
+
+-RH_API_HOST = "https://access.redhat.com"
++RH_API_HOST = "https://api.access.redhat.com"
+ RH_SFTP_HOST = "sftp://sftp.access.redhat.com"
+
+
+@@ -287,12 +287,12 @@ def upload_sftp(self):
+ " for obtaining SFTP auth token.")
+ _token = None
+ _user = None
++ url = RH_API_HOST + '/support/v2/sftp/token'
+ # we have a username and password, but we need to reset the password
+ # to be the token returned from the auth endpoint
+ if self.get_upload_user() and self.get_upload_password():
+- url = RH_API_HOST + '/hydra/rest/v1/sftp/token'
+ auth = self.get_upload_https_auth()
+- ret = requests.get(url, auth=auth, timeout=10)
++ ret = requests.post(url, auth=auth, timeout=10)
+ if ret.status_code == 200:
+ # credentials are valid
+ _user = self.get_upload_user()
+@@ -302,8 +302,8 @@ def upload_sftp(self):
+ "credentials. Will try anonymous.")
+ # we either do not have a username or password/token, or both
+ if not _token:
+- aurl = RH_API_HOST + '/hydra/rest/v1/sftp/token?isAnonymous=true'
+- anon = requests.get(aurl, timeout=10)
++ adata = {"isAnonymous": True}
++ anon = requests.post(url, data=json.dumps(adata), timeout=10)
+ if anon.status_code == 200:
+ resp = json.loads(anon.text)
+ _user = resp['username']
+
+From 267da2156ec61f526dd28e760ff6528408a76c3f Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 22 Nov 2021 15:22:32 +0100
+Subject: [PATCH 3/3] [policies] Deal 200 return code as success
+
+Return code 200 of POST method request must be dealt as success.
+
+Newly required due to the SFTP API change using POST.
+
+Related to: #2764
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/__init__.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index 0906fa779..6f257fdce 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -551,7 +551,7 @@ def upload_https(self):
+ r = self._upload_https_put(arc, verify)
+ else:
+ r = self._upload_https_post(arc, verify)
+- if r.status_code != 201:
++ if r.status_code != 200 and r.status_code != 201:
+ if r.status_code == 401:
+ raise Exception(
+ "Authentication failed: invalid user credentials"
+From 8da1b14246226792c160dd04e5c7c75dd4e8d44b Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 22 Nov 2021 10:44:09 +0100
+Subject: [PATCH] [collect] fix moved get_upload_url under Policy class
+
+SoSCollector does not further declare get_upload_url method
+as that was moved under Policy class(es).
+
+Resolves: #2766
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/collector/__init__.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index 50183e873..42a7731d6 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -1219,7 +1219,7 @@ this utility or remote systems that it c
+ msg = 'No sosreports were collected, nothing to archive...'
+ self.exit(msg, 1)
+
+- if self.opts.upload and self.get_upload_url():
++ if self.opts.upload and self.policy.get_upload_url():
+ try:
+ self.policy.upload_archive(arc_name)
+ self.ui_log.info("Uploaded archive successfully")
+From abb2fc65bd14760021c61699ad3113cab3bd4c64 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Tue, 30 Nov 2021 11:37:02 +0100
+Subject: [PATCH 1/2] [redhat] Fix broken URI to upload to customer portal
+
+Revert back the unwanted change in URI of uploading tarball to the
+Red Hat Customer portal.
+
+Related: #2772
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/redhat.py | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index e4e2b883..eb442407 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -250,7 +250,7 @@ support representative.
+ elif self.commons['cmdlineopts'].upload_protocol == 'sftp':
+ return RH_SFTP_HOST
+ else:
+- rh_case_api = "/hydra/rest/cases/%s/attachments"
++ rh_case_api = "/support/v1/cases/%s/attachments"
+ return RH_API_HOST + rh_case_api % self.case_id
+
+ def _get_upload_headers(self):
+--
+2.31.1
+
+
+From ea4f9e88a412c80a4791396e1bb78ac1e24ece14 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Tue, 30 Nov 2021 13:00:26 +0100
+Subject: [PATCH 2/2] [policy] Add error message when FTP upload write failure
+
+When (S)FTP upload fails to write the destination file,
+our "expect" code should detect it sooner than after timeout happens
+and write appropriate error message.
+
+Resolves: #2772
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/policies/distros/__init__.py | 5 ++++-
+ 1 file changed, 4 insertions(+), 1 deletion(-)
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index 6f257fdc..7bdc81b8 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -473,7 +473,8 @@ class LinuxPolicy(Policy):
+ put_expects = [
+ u'100%',
+ pexpect.TIMEOUT,
+- pexpect.EOF
++ pexpect.EOF,
++ u'No such file or directory'
+ ]
+
+ put_success = ret.expect(put_expects, timeout=180)
+@@ -485,6 +486,8 @@ class LinuxPolicy(Policy):
+ raise Exception("Timeout expired while uploading")
+ elif put_success == 2:
+ raise Exception("Unknown error during upload: %s" % ret.before)
++ elif put_success == 3:
++ raise Exception("Unable to write archive to destination")
+ else:
+ raise Exception("Unexpected response from server: %s" % ret.before)
+
+--
+2.31.1
+
diff --git a/SOURCES/sos-bz2031777-rhui-logs.patch b/SOURCES/sos-bz2031777-rhui-logs.patch
new file mode 100644
index 0000000..dcfbc89
--- /dev/null
+++ b/SOURCES/sos-bz2031777-rhui-logs.patch
@@ -0,0 +1,24 @@
+From aa2887f71c779448b22e4de67ae68dbaf218b7b9 Mon Sep 17 00:00:00 2001
+From: Taft Sanders <taftsanders@gmail.com>
+Date: Fri, 10 Dec 2021 09:34:59 -0500
+Subject: [PATCH] [rhui] New log folder
+
+Included new log folder per Bugzilla 2030741
+
+Signed-off-by: Taft Sanders <taftsanders@gmail.com>
+---
+ sos/report/plugins/rhui.py | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/sos/report/plugins/rhui.py b/sos/report/plugins/rhui.py
+index 52065fb44..add024613 100644
+--- a/sos/report/plugins/rhui.py
++++ b/sos/report/plugins/rhui.py
+@@ -27,6 +27,7 @@ def setup(self):
+ "/var/log/rhui-subscription-sync.log",
+ "/var/cache/rhui/*",
+ "/root/.rhui/*",
++ "/var/log/rhui/*",
+ ])
+ # skip collecting certificate keys
+ self.add_forbidden_path("/etc/pki/rhui/**/*.key", recursive=True)
diff --git a/SOURCES/sos-bz2034001-nvidia-GPU-info.patch b/SOURCES/sos-bz2034001-nvidia-GPU-info.patch
new file mode 100644
index 0000000..30fbb53
--- /dev/null
+++ b/SOURCES/sos-bz2034001-nvidia-GPU-info.patch
@@ -0,0 +1,46 @@
+From f2cc67750f55a71edff0c527a1bfc14fde8132c3 Mon Sep 17 00:00:00 2001
+From: Mamatha Inamdar <mamatha4@linux.vnet.ibm.com>
+Date: Mon, 8 Nov 2021 10:50:03 +0530
+Subject: [PATCH] [nvidia]:Patch to update nvidia plugin for GPU info
+
+This patch is to update nvidia plugin to collect
+logs for Nvidia GPUs
+
+Signed-off-by: Mamatha Inamdar <mamatha4@linux.vnet.ibm.com>
+Reported-by: Borislav Stoymirski <borislav.stoymirski@bg.ibm.com>
+Reported-by: Yesenia Jimenez <yesenia@us.ibm.com>
+---
+ sos/report/plugins/nvidia.py | 15 +++++++++++++--
+ 1 file changed, 13 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/nvidia.py b/sos/report/plugins/nvidia.py
+index 09aaf586b..9e21b478e 100644
+--- a/sos/report/plugins/nvidia.py
++++ b/sos/report/plugins/nvidia.py
+@@ -23,13 +23,24 @@ def setup(self):
+ '--list-gpus',
+ '-q -d PERFORMANCE',
+ '-q -d SUPPORTED_CLOCKS',
+- '-q -d PAGE_RETIREMENT'
++ '-q -d PAGE_RETIREMENT',
++ '-q',
++ '-q -d ECC',
++ 'nvlink -s',
++ 'nvlink -e'
+ ]
+
+ self.add_cmd_output(["nvidia-smi %s" % cmd for cmd in subcmds])
+
+ query = ('gpu_name,gpu_bus_id,vbios_version,temperature.gpu,'
+- 'utilization.gpu,memory.total,memory.free,memory.used')
++ 'utilization.gpu,memory.total,memory.free,memory.used,'
++ 'clocks.applications.graphics,clocks.applications.memory')
++ querypages = ('timestamp,gpu_bus_id,gpu_serial,gpu_uuid,'
++ 'retired_pages.address,retired_pages.cause')
+ self.add_cmd_output("nvidia-smi --query-gpu=%s --format=csv" % query)
++ self.add_cmd_output(
++ "nvidia-smi --query-retired-pages=%s --format=csv" % querypages
++ )
++ self.add_journal(boot=0, identifier='nvidia-persistenced')
+
+ # vim: set et ts=4 sw=4 :
diff --git a/SOURCES/sos-bz2037350-ocp-backports.patch b/SOURCES/sos-bz2037350-ocp-backports.patch
new file mode 100644
index 0000000..3e53e93
--- /dev/null
+++ b/SOURCES/sos-bz2037350-ocp-backports.patch
@@ -0,0 +1,5145 @@
+From 676dfca09d9c783311a51a1c53fa0f7ecd95bd28 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Fri, 10 Sep 2021 13:38:19 -0400
+Subject: [PATCH] [collect] Abstract transport protocol from SoSNode
+
+Since its addition to sos, collect has assumed the use of a system
+installation of SSH in order to connect to the nodes identified for
+collection. However, there may be use cases and desires to use other
+transport protocols.
+
+As such, provide an abstraction for these protocols in the form of the
+new `RemoteTransport` class that `SoSNode` will now leverage. So far an
+abstraction for the currently used SSH ControlPersist function is
+provided, along with a psuedo abstraction for local execution so that
+SoSNode does not directly need to make more "if local then foo" checks
+than are absolutely necessary.
+
+Related: #2668
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ setup.py | 4 +-
+ sos/collector/__init__.py | 54 +--
+ sos/collector/clusters/__init__.py | 4 +-
+ sos/collector/clusters/jbon.py | 2 +
+ sos/collector/clusters/kubernetes.py | 4 +-
+ sos/collector/clusters/ocp.py | 6 +-
+ sos/collector/clusters/ovirt.py | 10 +-
+ sos/collector/clusters/pacemaker.py | 8 +-
+ sos/collector/clusters/satellite.py | 4 +-
+ sos/collector/sosnode.py | 388 +++++---------------
+ sos/collector/transports/__init__.py | 317 ++++++++++++++++
+ sos/collector/transports/control_persist.py | 199 ++++++++++
+ sos/collector/transports/local.py | 49 +++
+ 13 files changed, 705 insertions(+), 344 deletions(-)
+ create mode 100644 sos/collector/transports/__init__.py
+ create mode 100644 sos/collector/transports/control_persist.py
+ create mode 100644 sos/collector/transports/local.py
+
+diff --git a/setup.py b/setup.py
+index 7653b59d..25e87a71 100644
+--- a/setup.py
++++ b/setup.py
+@@ -101,8 +101,8 @@ setup(
+ 'sos.policies.distros', 'sos.policies.runtimes',
+ 'sos.policies.package_managers', 'sos.policies.init_systems',
+ 'sos.report', 'sos.report.plugins', 'sos.collector',
+- 'sos.collector.clusters', 'sos.cleaner', 'sos.cleaner.mappings',
+- 'sos.cleaner.parsers', 'sos.cleaner.archives'
++ 'sos.collector.clusters', 'sos.collector.transports', 'sos.cleaner',
++ 'sos.cleaner.mappings', 'sos.cleaner.parsers', 'sos.cleaner.archives'
+ ],
+ cmdclass=cmdclass,
+ command_options=command_options,
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index b2a07f37..da912655 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -17,7 +17,6 @@ import re
+ import string
+ import socket
+ import shutil
+-import subprocess
+ import sys
+
+ from datetime import datetime
+@@ -28,7 +27,6 @@ from pipes import quote
+ from textwrap import fill
+ from sos.cleaner import SoSCleaner
+ from sos.collector.sosnode import SosNode
+-from sos.collector.exceptions import ControlPersistUnsupportedException
+ from sos.options import ClusterOption
+ from sos.component import SoSComponent
+ from sos import __version__
+@@ -154,7 +152,6 @@ class SoSCollector(SoSComponent):
+ try:
+ self.parse_node_strings()
+ self.parse_cluster_options()
+- self._check_for_control_persist()
+ self.log_debug('Executing %s' % ' '.join(s for s in sys.argv))
+ self.log_debug("Found cluster profiles: %s"
+ % self.clusters.keys())
+@@ -437,33 +434,6 @@ class SoSCollector(SoSComponent):
+ action='extend',
+ help='List of usernames to obfuscate')
+
+- def _check_for_control_persist(self):
+- """Checks to see if the local system supported SSH ControlPersist.
+-
+- ControlPersist allows OpenSSH to keep a single open connection to a
+- remote host rather than building a new session each time. This is the
+- same feature that Ansible uses in place of paramiko, which we have a
+- need to drop in sos-collector.
+-
+- This check relies on feedback from the ssh binary. The command being
+- run should always generate stderr output, but depending on what that
+- output reads we can determine if ControlPersist is supported or not.
+-
+- For our purposes, a host that does not support ControlPersist is not
+- able to run sos-collector.
+-
+- Returns
+- True if ControlPersist is supported, else raise Exception.
+- """
+- ssh_cmd = ['ssh', '-o', 'ControlPersist']
+- cmd = subprocess.Popen(ssh_cmd, stdout=subprocess.PIPE,
+- stderr=subprocess.PIPE)
+- out, err = cmd.communicate()
+- err = err.decode('utf-8')
+- if 'Bad configuration option' in err or 'Usage:' in err:
+- raise ControlPersistUnsupportedException
+- return True
+-
+ def exit(self, msg, error=1):
+ """Used to safely terminate if sos-collector encounters an error"""
+ self.log_error(msg)
+@@ -455,7 +455,7 @@ class SoSCollector(SoSComponent):
+ 'cmdlineopts': self.opts,
+ 'need_sudo': True if self.opts.ssh_user != 'root' else False,
+ 'tmpdir': self.tmpdir,
+- 'hostlen': len(self.opts.master) or len(self.hostname),
++ 'hostlen': max(len(self.opts.primary), len(self.hostname)),
+ 'policy': self.policy
+ }
+
+@@ -1020,9 +1020,10 @@ class SoSCollector(SoSComponent):
+ self.node_list.append(self.hostname)
+ self.reduce_node_list()
+ try:
+- self.commons['hostlen'] = len(max(self.node_list, key=len))
++ _node_max = len(max(self.node_list, key=len))
++ self.commons['hostlen'] = max(_node_max, self.commons['hostlen'])
+ except (TypeError, ValueError):
+- self.commons['hostlen'] = len(self.opts.master)
++ pass
+
+ def _connect_to_node(self, node):
+ """Try to connect to the node, and if we can add to the client list to
+@@ -1068,7 +1039,7 @@ class SoSCollector(SoSComponent):
+ client.set_node_manifest(getattr(self.collect_md.nodes,
+ node[0]))
+ else:
+- client.close_ssh_session()
++ client.disconnect()
+ except Exception:
+ pass
+
+@@ -1077,12 +1048,11 @@ class SoSCollector(SoSComponent):
+ provided on the command line
+ """
+ disclaimer = ("""\
+-This utility is used to collect sosreports from multiple \
+-nodes simultaneously. It uses OpenSSH's ControlPersist feature \
+-to connect to nodes and run commands remotely. If your system \
+-installation of OpenSSH is older than 5.6, please upgrade.
++This utility is used to collect sos reports from multiple \
++nodes simultaneously. Remote connections are made and/or maintained \
++to those nodes via well-known transport protocols such as SSH.
+
+-An archive of sosreport tarballs collected from the nodes will be \
++An archive of sos report tarballs collected from the nodes will be \
+ generated in %s and may be provided to an appropriate support representative.
+
+ The generated archive may contain data considered sensitive \
+@@ -1230,10 +1200,10 @@ this utility or remote systems that it connects to.
+ self.log_error("Error running sosreport: %s" % err)
+
+ def close_all_connections(self):
+- """Close all ssh sessions for nodes"""
++ """Close all sessions for nodes"""
+ for client in self.client_list:
+- self.log_debug('Closing SSH connection to %s' % client.address)
+- client.close_ssh_session()
++ self.log_debug('Closing connection to %s' % client.address)
++ client.disconnect()
+
+ def create_cluster_archive(self):
+ """Calls for creation of tar archive then cleans up the temporary
+diff --git a/sos/collector/clusters/__init__.py b/sos/collector/clusters/__init__.py
+index 2b5d7018..64ac2a44 100644
+--- a/sos/collector/clusters/__init__.py
++++ b/sos/collector/clusters/__init__.py
+@@ -188,8 +188,8 @@ class Cluster():
+ :rtype: ``dict``
+ """
+ res = self.master.run_command(cmd, get_pty=True, need_root=need_root)
+- if res['stdout']:
+- res['stdout'] = res['stdout'].replace('Password:', '')
++ if res['output']:
++ res['output'] = res['output'].replace('Password:', '')
+ return res
+
+ def setup(self):
+diff --git a/sos/collector/clusters/jbon.py b/sos/collector/clusters/jbon.py
+index 488fbd16..8f083ac6 100644
+--- a/sos/collector/clusters/jbon.py
++++ b/sos/collector/clusters/jbon.py
+@@ -28,3 +28,5 @@ class jbon(Cluster):
+ # This should never be called, but as insurance explicitly never
+ # allow this to be enabled via the determine_cluster() path
+ return False
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/clusters/kubernetes.py b/sos/collector/clusters/kubernetes.py
+index cdbf8861..99f788dc 100644
+--- a/sos/collector/clusters/kubernetes.py
++++ b/sos/collector/clusters/kubernetes.py
+@@ -34,7 +34,7 @@ class kubernetes(Cluster):
+ if res['status'] == 0:
+ nodes = []
+ roles = [x for x in self.get_option('role').split(',') if x]
+- for nodeln in res['stdout'].splitlines()[1:]:
++ for nodeln in res['output'].splitlines()[1:]:
+ node = nodeln.split()
+ if not roles:
+ nodes.append(node[0])
+@@ -44,3 +44,5 @@ class kubernetes(Cluster):
+ return nodes
+ else:
+ raise Exception('Node enumeration did not return usable output')
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index 5479417d..ad97587f 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -93,7 +93,7 @@ class ocp(Cluster):
+ res = self.exec_master_cmd(self.fmt_oc_cmd(cmd))
+ if res['status'] == 0:
+ roles = [r for r in self.get_option('role').split(':')]
+- self.node_dict = self._build_dict(res['stdout'].splitlines())
++ self.node_dict = self._build_dict(res['output'].splitlines())
+ for node in self.node_dict:
+ if roles:
+ for role in roles:
+@@ -103,7 +103,7 @@ class ocp(Cluster):
+ nodes.append(node)
+ else:
+ msg = "'oc' command failed"
+- if 'Missing or incomplete' in res['stdout']:
++ if 'Missing or incomplete' in res['output']:
+ msg = ("'oc' failed due to missing kubeconfig on master node."
+ " Specify one via '-c ocp.kubeconfig=<path>'")
+ raise Exception(msg)
+@@ -168,3 +168,5 @@ class ocp(Cluster):
+ def set_node_options(self, node):
+ # don't attempt OC API collections on non-primary nodes
+ node.plugin_options.append('openshift.no-oc=on')
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/clusters/ovirt.py b/sos/collector/clusters/ovirt.py
+index 079a122e..bd2d0c74 100644
+--- a/sos/collector/clusters/ovirt.py
++++ b/sos/collector/clusters/ovirt.py
+@@ -98,7 +98,7 @@ class ovirt(Cluster):
+ return []
+ res = self._run_db_query(self.dbquery)
+ if res['status'] == 0:
+- nodes = res['stdout'].splitlines()[2:-1]
++ nodes = res['output'].splitlines()[2:-1]
+ return [n.split('(')[0].strip() for n in nodes]
+ else:
+ raise Exception('database query failed, return code: %s'
+@@ -114,7 +114,7 @@ class ovirt(Cluster):
+ engconf = '/etc/ovirt-engine/engine.conf.d/10-setup-database.conf'
+ res = self.exec_primary_cmd('cat %s' % engconf, need_root=True)
+ if res['status'] == 0:
+- config = res['stdout'].splitlines()
++ config = res['output'].splitlines()
+ for line in config:
+ try:
+ k = str(line.split('=')[0])
+@@ -141,7 +141,7 @@ class ovirt(Cluster):
+ '--batch -o postgresql {}'
+ ).format(self.conf['ENGINE_DB_PASSWORD'], sos_opt)
+ db_sos = self.exec_primary_cmd(cmd, need_root=True)
+- for line in db_sos['stdout'].splitlines():
++ for line in db_sos['output'].splitlines():
+ if fnmatch.fnmatch(line, '*sosreport-*tar*'):
+ _pg_dump = line.strip()
+ self.master.manifest.add_field('postgresql_dump',
+@@ -180,5 +180,7 @@ class rhhi_virt(rhv):
+ ret = self._run_db_query('SELECT count(server_id) FROM gluster_server')
+ if ret['status'] == 0:
+ # if there are any entries in this table, RHHI-V is in use
+- return ret['stdout'].splitlines()[2].strip() != '0'
++ return ret['output'].splitlines()[2].strip() != '0'
+ return False
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/clusters/pacemaker.py b/sos/collector/clusters/pacemaker.py
+index 034f3f3e..55024314 100644
+--- a/sos/collector/clusters/pacemaker.py
++++ b/sos/collector/clusters/pacemaker.py
+@@ -27,7 +27,7 @@ class pacemaker(Cluster):
+ self.log_error('Cluster status could not be determined. Is the '
+ 'cluster running on this node?')
+ return []
+- if 'node names do not match' in self.res['stdout']:
++ if 'node names do not match' in self.res['output']:
+ self.log_warn('Warning: node name mismatch reported. Attempts to '
+ 'connect to some nodes may fail.\n')
+ return self.parse_pcs_output()
+@@ -41,17 +41,19 @@ class pacemaker(Cluster):
+ return nodes
+
+ def get_online_nodes(self):
+- for line in self.res['stdout'].splitlines():
++ for line in self.res['output'].splitlines():
+ if line.startswith('Online:'):
+ nodes = line.split('[')[1].split(']')[0]
+ return [n for n in nodes.split(' ') if n]
+
+ def get_offline_nodes(self):
+ offline = []
+- for line in self.res['stdout'].splitlines():
++ for line in self.res['output'].splitlines():
+ if line.startswith('Node') and line.endswith('(offline)'):
+ offline.append(line.split()[1].replace(':', ''))
+ if line.startswith('OFFLINE:'):
+ nodes = line.split('[')[1].split(']')[0]
+ offline.extend([n for n in nodes.split(' ') if n])
+ return offline
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/clusters/satellite.py b/sos/collector/clusters/satellite.py
+index e123c8a3..7c21e553 100644
+--- a/sos/collector/clusters/satellite.py
++++ b/sos/collector/clusters/satellite.py
+@@ -28,7 +28,7 @@ class satellite(Cluster):
+ res = self.exec_primary_cmd(cmd, need_root=True)
+ if res['status'] == 0:
+ nodes = [
+- n.strip() for n in res['stdout'].splitlines()
++ n.strip() for n in res['output'].splitlines()
+ if 'could not change directory' not in n
+ ]
+ return nodes
+@@ -38,3 +38,5 @@ class satellite(Cluster):
+ if node.address == self.master.address:
+ return 'satellite'
+ return 'capsule'
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 4b1ee109..f79bd5ff 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -12,22 +12,16 @@ import fnmatch
+ import inspect
+ import logging
+ import os
+-import pexpect
+ import re
+-import shutil
+
+ from distutils.version import LooseVersion
+ from pipes import quote
+ from sos.policies import load
+ from sos.policies.init_systems import InitSystem
+-from sos.collector.exceptions import (InvalidPasswordException,
+- TimeoutPasswordAuthException,
+- PasswordRequestException,
+- AuthPermissionDeniedException,
++from sos.collector.transports.control_persist import SSHControlPersist
++from sos.collector.transports.local import LocalTransport
++from sos.collector.exceptions import (CommandTimeoutException,
+ ConnectionException,
+- CommandTimeoutException,
+- ConnectionTimeoutException,
+- ControlSocketMissingException,
+ UnsupportedHostException)
+
+
+@@ -61,34 +61,25 @@ class SosNode():
+ 'sos_cmd': commons['sos_cmd']
+ }
+ self.sos_bin = 'sosreport'
+- filt = ['localhost', '127.0.0.1']
+ self.soslog = logging.getLogger('sos')
+ self.ui_log = logging.getLogger('sos_ui')
+- self.control_path = ("%s/.sos-collector-%s"
+- % (self.tmpdir, self.address))
+- self.ssh_cmd = self._create_ssh_command()
+- if self.address not in filt:
+- try:
+- self.connected = self._create_ssh_session()
+- except Exception as err:
+- self.log_error('Unable to open SSH session: %s' % err)
+- raise
+- else:
+- self.connected = True
+- self.local = True
+- self.need_sudo = os.getuid() != 0
++ self._transport = self._load_remote_transport(commons)
++ try:
++ self._transport.connect(self._password)
++ except Exception as err:
++ self.log_error('Unable to open remote session: %s' % err)
++ raise
+ # load the host policy now, even if we don't want to load further
+ # host information. This is necessary if we're running locally on the
+ # cluster master but do not want a local report as we still need to do
+ # package checks in that instance
+ self.host = self.determine_host_policy()
+- self.get_hostname()
++ self.hostname = self._transport.hostname
+ if self.local and self.opts.no_local:
+ load_facts = False
+ if self.connected and load_facts:
+ if not self.host:
+- self.connected = False
+- self.close_ssh_session()
++ self._transport.disconnect()
+ return None
+ if self.local:
+ if self.check_in_container():
+@@ -103,11 +88,26 @@ class SosNode():
+ self.create_sos_container()
+ self._load_sos_info()
+
+- def _create_ssh_command(self):
+- """Build the complete ssh command for this node"""
+- cmd = "ssh -oControlPath=%s " % self.control_path
+- cmd += "%s@%s " % (self.opts.ssh_user, self.address)
+- return cmd
++ @property
++ def connected(self):
++ if self._transport:
++ return self._transport.connected
++ # if no transport, we're running locally
++ return True
++
++ def disconnect(self):
++ """Wrapper to close the remote session via our transport agent
++ """
++ self._transport.disconnect()
++
++ def _load_remote_transport(self, commons):
++ """Determine the type of remote transport to load for this node, then
++ return an instantiated instance of that transport
++ """
++ if self.address in ['localhost', '127.0.0.1']:
++ self.local = True
++ return LocalTransport(self.address, commons)
++ return SSHControlPersist(self.address, commons)
+
+ def _fmt_msg(self, msg):
+ return '{:<{}} : {}'.format(self._hostname, self.hostlen + 1, msg)
+@@ -135,6 +135,7 @@ class SosNode():
+ self.manifest.add_field('policy', self.host.distro)
+ self.manifest.add_field('sos_version', self.sos_info['version'])
+ self.manifest.add_field('final_sos_command', '')
++ self.manifest.add_field('transport', self._transport.name)
+
+ def check_in_container(self):
+ """
+@@ -160,13 +161,13 @@ class SosNode():
+ res = self.run_command(cmd, need_root=True)
+ if res['status'] in [0, 125]:
+ if res['status'] == 125:
+- if 'unable to retrieve auth token' in res['stdout']:
++ if 'unable to retrieve auth token' in res['output']:
+ self.log_error(
+ "Could not pull image. Provide either a username "
+ "and password or authfile"
+ )
+ raise Exception
+- elif 'unknown: Not found' in res['stdout']:
++ elif 'unknown: Not found' in res['output']:
+ self.log_error('Specified image not found on registry')
+ raise Exception
+ # 'name exists' with code 125 means the container was
+@@ -181,11 +182,11 @@ class SosNode():
+ return True
+ else:
+ self.log_error("Could not start container after create: %s"
+- % ret['stdout'])
++ % ret['output'])
+ raise Exception
+ else:
+ self.log_error("Could not create container on host: %s"
+- % res['stdout'])
++ % res['output'])
+ raise Exception
+
+ def get_container_auth(self):
+@@ -204,18 +205,11 @@ class SosNode():
+
+ def file_exists(self, fname, need_root=False):
+ """Checks for the presence of fname on the remote node"""
+- if not self.local:
+- try:
+- res = self.run_command("stat %s" % fname, need_root=need_root)
+- return res['status'] == 0
+- except Exception:
+- return False
+- else:
+- try:
+- os.stat(fname)
+- return True
+- except Exception:
+- return False
++ try:
++ res = self.run_command("stat %s" % fname, need_root=need_root)
++ return res['status'] == 0
++ except Exception:
++ return False
+
+ @property
+ def _hostname(self):
+@@ -223,18 +217,6 @@ class SosNode():
+ return self.hostname
+ return self.address
+
+- @property
+- def control_socket_exists(self):
+- """Check if the SSH control socket exists
+-
+- The control socket is automatically removed by the SSH daemon in the
+- event that the last connection to the node was greater than the timeout
+- set by the ControlPersist option. This can happen for us if we are
+- collecting from a large number of nodes, and the timeout expires before
+- we start collection.
+- """
+- return os.path.exists(self.control_path)
+-
+ def _sanitize_log_msg(self, msg):
+ """Attempts to obfuscate sensitive information in log messages such as
+ passwords"""
+@@ -264,12 +246,6 @@ class SosNode():
+ msg = '[%s:%s] %s' % (self._hostname, caller, msg)
+ self.soslog.debug(msg)
+
+- def get_hostname(self):
+- """Get the node's hostname"""
+- sout = self.run_command('hostname')
+- self.hostname = sout['stdout'].strip()
+- self.log_info('Hostname set to %s' % self.hostname)
+-
+ def _format_cmd(self, cmd):
+ """If we need to provide a sudo or root password to a command, then
+ here we prefix the command with the correct bits
+@@ -280,19 +256,6 @@ class SosNode():
+ return "sudo -S %s" % cmd
+ return cmd
+
+- def _fmt_output(self, output=None, rc=0):
+- """Formats the returned output from a command into a dict"""
+- if rc == 0:
+- stdout = output
+- stderr = ''
+- else:
+- stdout = ''
+- stderr = output
+- res = {'status': rc,
+- 'stdout': stdout,
+- 'stderr': stderr}
+- return res
+-
+ def _load_sos_info(self):
+ """Queries the node for information about the installed version of sos
+ """
+@@ -306,7 +269,7 @@ class SosNode():
+ pkgs = self.run_command(self.host.container_version_command,
+ use_container=True, need_root=True)
+ if pkgs['status'] == 0:
+- ver = pkgs['stdout'].strip().split('-')[1]
++ ver = pkgs['output'].strip().split('-')[1]
+ if ver:
+ self.sos_info['version'] = ver
+ else:
+@@ -321,18 +284,21 @@ class SosNode():
+ self.log_error('sos is not installed on this node')
+ self.connected = False
+ return False
+- cmd = 'sosreport -l'
++ # sos-4.0 changes the binary
++ if self.check_sos_version('4.0'):
++ self.sos_bin = 'sos report'
++ cmd = "%s -l" % self.sos_bin
+ sosinfo = self.run_command(cmd, use_container=True, need_root=True)
+ if sosinfo['status'] == 0:
+- self._load_sos_plugins(sosinfo['stdout'])
++ self._load_sos_plugins(sosinfo['output'])
+ if self.check_sos_version('3.6'):
+ self._load_sos_presets()
+
+ def _load_sos_presets(self):
+- cmd = 'sosreport --list-presets'
++ cmd = '%s --list-presets' % self.sos_bin
+ res = self.run_command(cmd, use_container=True, need_root=True)
+ if res['status'] == 0:
+- for line in res['stdout'].splitlines():
++ for line in res['output'].splitlines():
+ if line.strip().startswith('name:'):
+ pname = line.split('name:')[1].strip()
+ self.sos_info['presets'].append(pname)
+@@ -372,21 +338,7 @@ class SosNode():
+ """Reads the specified file and returns the contents"""
+ try:
+ self.log_info("Reading file %s" % to_read)
+- if not self.local:
+- res = self.run_command("cat %s" % to_read, timeout=5)
+- if res['status'] == 0:
+- return res['stdout']
+- else:
+- if 'No such file' in res['stdout']:
+- self.log_debug("File %s does not exist on node"
+- % to_read)
+- else:
+- self.log_error("Error reading %s: %s" %
+- (to_read, res['stdout'].split(':')[1:]))
+- return ''
+- else:
+- with open(to_read, 'r') as rfile:
+- return rfile.read()
++ return self._transport.read_file(to_read)
+ except Exception as err:
+ self.log_error("Exception while reading %s: %s" % (to_read, err))
+ return ''
+@@ -400,7 +352,8 @@ class SosNode():
+ % self.commons['policy'].distro)
+ return self.commons['policy']
+ host = load(cache={}, sysroot=self.opts.sysroot, init=InitSystem(),
+- probe_runtime=True, remote_exec=self.ssh_cmd,
++ probe_runtime=True,
++ remote_exec=self._transport.remote_exec,
+ remote_check=self.read_file('/etc/os-release'))
+ if host:
+ self.log_info("loaded policy %s for host" % host.distro)
+@@ -422,7 +375,7 @@ class SosNode():
+ return self.host.package_manager.pkg_by_name(pkg) is not None
+
+ def run_command(self, cmd, timeout=180, get_pty=False, need_root=False,
+- force_local=False, use_container=False, env=None):
++ use_container=False, env=None):
+ """Runs a given cmd, either via the SSH session or locally
+
+ Arguments:
+@@ -433,58 +386,37 @@ class SosNode():
+ need_root - if a command requires root privileges, setting this to
+ True tells sos-collector to format the command with
+ sudo or su - as appropriate and to input the password
+- force_local - force a command to run locally. Mainly used for scp.
+ use_container - Run this command in a container *IF* the host is
+ containerized
+ """
+- if not self.control_socket_exists and not self.local:
+- self.log_debug('Control socket does not exist, attempting to '
+- 're-create')
++ if not self.connected and not self.local:
++ self.log_debug('Node is disconnected, attempting to reconnect')
+ try:
+- _sock = self._create_ssh_session()
+- if not _sock:
+- self.log_debug('Failed to re-create control socket')
+- raise ControlSocketMissingException
++ reconnected = self._transport.reconnect(self._password)
++ if not reconnected:
++ self.log_debug('Failed to reconnect to node')
++ raise ConnectionException
+ except Exception as err:
+- self.log_error('Cannot run command: control socket does not '
+- 'exist')
+- self.log_debug("Error while trying to create new SSH control "
+- "socket: %s" % err)
++ self.log_debug("Error while trying to reconnect: %s" % err)
+ raise
+ if use_container and self.host.containerized:
+ cmd = self.host.format_container_command(cmd)
+ if need_root:
+- get_pty = True
+ cmd = self._format_cmd(cmd)
+- self.log_debug('Running command %s' % cmd)
++
+ if 'atomic' in cmd:
+ get_pty = True
+- if not self.local and not force_local:
+- cmd = "%s %s" % (self.ssh_cmd, quote(cmd))
+- else:
+- if get_pty:
+- cmd = "/bin/bash -c %s" % quote(cmd)
++
++ if get_pty:
++ cmd = "/bin/bash -c %s" % quote(cmd)
++
+ if env:
+ _cmd_env = self.env_vars
+ env = _cmd_env.update(env)
+- res = pexpect.spawn(cmd, encoding='utf-8', env=env)
+- if need_root:
+- if self.need_sudo:
+- res.sendline(self.opts.sudo_pw)
+- if self.opts.become_root:
+- res.sendline(self.opts.root_password)
+- output = res.expect([pexpect.EOF, pexpect.TIMEOUT],
+- timeout=timeout)
+- if output == 0:
+- out = res.before
+- res.close()
+- rc = res.exitstatus
+- return {'status': rc, 'stdout': out}
+- elif output == 1:
+- raise CommandTimeoutException(cmd)
++ return self._transport.run_command(cmd, timeout, need_root, env)
+
+ def sosreport(self):
+- """Run a sosreport on the node, then collect it"""
++ """Run an sos report on the node, then collect it"""
+ try:
+ path = self.execute_sos_command()
+ if path:
+@@ -497,109 +429,6 @@ class SosNode():
+ pass
+ self.cleanup()
+
+- def _create_ssh_session(self):
+- """
+- Using ControlPersist, create the initial connection to the node.
+-
+- This will generate an OpenSSH ControlPersist socket within the tmp
+- directory created or specified for sos-collector to use.
+-
+- At most, we will wait 30 seconds for a connection. This involves a 15
+- second wait for the initial connection attempt, and a subsequent 15
+- second wait for a response when we supply a password.
+-
+- Since we connect to nodes in parallel (using the --threads value), this
+- means that the time between 'Connecting to nodes...' and 'Beginning
+- collection of sosreports' that users see can be up to an amount of time
+- equal to 30*(num_nodes/threads) seconds.
+-
+- Returns
+- True if session is successfully opened, else raise Exception
+- """
+- # Don't use self.ssh_cmd here as we need to add a few additional
+- # parameters to establish the initial connection
+- self.log_info('Opening SSH session to create control socket')
+- connected = False
+- ssh_key = ''
+- ssh_port = ''
+- if self.opts.ssh_port != 22:
+- ssh_port = "-p%s " % self.opts.ssh_port
+- if self.opts.ssh_key:
+- ssh_key = "-i%s" % self.opts.ssh_key
+- cmd = ("ssh %s %s -oControlPersist=600 -oControlMaster=auto "
+- "-oStrictHostKeyChecking=no -oControlPath=%s %s@%s "
+- "\"echo Connected\"" % (ssh_key,
+- ssh_port,
+- self.control_path,
+- self.opts.ssh_user,
+- self.address))
+- res = pexpect.spawn(cmd, encoding='utf-8')
+-
+- connect_expects = [
+- u'Connected',
+- u'password:',
+- u'.*Permission denied.*',
+- u'.* port .*: No route to host',
+- u'.*Could not resolve hostname.*',
+- pexpect.TIMEOUT
+- ]
+-
+- index = res.expect(connect_expects, timeout=15)
+-
+- if index == 0:
+- connected = True
+- elif index == 1:
+- if self._password:
+- pass_expects = [
+- u'Connected',
+- u'Permission denied, please try again.',
+- pexpect.TIMEOUT
+- ]
+- res.sendline(self._password)
+- pass_index = res.expect(pass_expects, timeout=15)
+- if pass_index == 0:
+- connected = True
+- elif pass_index == 1:
+- # Note that we do not get an exitstatus here, so matching
+- # this line means an invalid password will be reported for
+- # both invalid passwords and invalid user names
+- raise InvalidPasswordException
+- elif pass_index == 2:
+- raise TimeoutPasswordAuthException
+- else:
+- raise PasswordRequestException
+- elif index == 2:
+- raise AuthPermissionDeniedException
+- elif index == 3:
+- raise ConnectionException(self.address, self.opts.ssh_port)
+- elif index == 4:
+- raise ConnectionException(self.address)
+- elif index == 5:
+- raise ConnectionTimeoutException
+- else:
+- raise Exception("Unknown error, client returned %s" % res.before)
+- if connected:
+- self.log_debug("Successfully created control socket at %s"
+- % self.control_path)
+- return True
+- return False
+-
+- def close_ssh_session(self):
+- """Remove the control socket to effectively terminate the session"""
+- if self.local:
+- return True
+- try:
+- res = self.run_command("rm -f %s" % self.control_path,
+- force_local=True)
+- if res['status'] == 0:
+- return True
+- self.log_error("Could not remove ControlPath %s: %s"
+- % (self.control_path, res['stdout']))
+- return False
+- except Exception as e:
+- self.log_error('Error closing SSH session: %s' % e)
+- return False
+-
+ def _preset_exists(self, preset):
+ """Verifies if the given preset exists on the node"""
+ return preset in self.sos_info['presets']
+@@ -646,8 +475,8 @@ class SosNode():
+ self.cluster = cluster
+
+ def update_cmd_from_cluster(self):
+- """This is used to modify the sosreport command run on the nodes.
+- By default, sosreport is run without any options, using this will
++ """This is used to modify the sos report command run on the nodes.
++ By default, sos report is run without any options, using this will
+ allow the profile to specify what plugins to run or not and what
+ options to use.
+
+@@ -727,10 +556,6 @@ class SosNode():
+ if self.opts.since:
+ sos_opts.append('--since=%s' % quote(self.opts.since))
+
+- # sos-4.0 changes the binary
+- if self.check_sos_version('4.0'):
+- self.sos_bin = 'sos report'
+-
+ if self.check_sos_version('4.1'):
+ if self.opts.skip_commands:
+ sos_opts.append(
+@@ -811,7 +636,7 @@ class SosNode():
+ self.manifest.add_field('final_sos_command', self.sos_cmd)
+
+ def determine_sos_label(self):
+- """Determine what, if any, label should be added to the sosreport"""
++ """Determine what, if any, label should be added to the sos report"""
+ label = ''
+ label += self.cluster.get_node_label(self)
+
+@@ -822,7 +647,7 @@ class SosNode():
+ if not label:
+ return None
+
+- self.log_debug('Label for sosreport set to %s' % label)
++ self.log_debug('Label for sos report set to %s' % label)
+ if self.check_sos_version('3.6'):
+ lcmd = '--label'
+ else:
+@@ -844,20 +669,20 @@ class SosNode():
+
+ def determine_sos_error(self, rc, stdout):
+ if rc == -1:
+- return 'sosreport process received SIGKILL on node'
++ return 'sos report process received SIGKILL on node'
+ if rc == 1:
+ if 'sudo' in stdout:
+ return 'sudo attempt failed'
+ if rc == 127:
+- return 'sosreport terminated unexpectedly. Check disk space'
++ return 'sos report terminated unexpectedly. Check disk space'
+ if len(stdout) > 0:
+ return stdout.split('\n')[0:1]
+ else:
+ return 'sos exited with code %s' % rc
+
+ def execute_sos_command(self):
+- """Run sosreport and capture the resulting file path"""
+- self.ui_msg('Generating sosreport...')
++ """Run sos report and capture the resulting file path"""
++ self.ui_msg('Generating sos report...')
+ try:
+ path = False
+ checksum = False
+@@ -867,7 +692,7 @@ class SosNode():
+ use_container=True,
+ env=self.sos_env_vars)
+ if res['status'] == 0:
+- for line in res['stdout'].splitlines():
++ for line in res['output'].splitlines():
+ if fnmatch.fnmatch(line, '*sosreport-*tar*'):
+ path = line.strip()
+ if line.startswith((" sha256\t", " md5\t")):
+@@ -884,44 +709,31 @@ class SosNode():
+ else:
+ self.manifest.add_field('checksum_type', 'unknown')
+ else:
+- err = self.determine_sos_error(res['status'], res['stdout'])
+- self.log_debug("Error running sosreport. rc = %s msg = %s"
+- % (res['status'], res['stdout'] or
+- res['stderr']))
++ err = self.determine_sos_error(res['status'], res['output'])
++ self.log_debug("Error running sos report. rc = %s msg = %s"
++ % (res['status'], res['output']))
+ raise Exception(err)
+ return path
+ except CommandTimeoutException:
+ self.log_error('Timeout exceeded')
+ raise
+ except Exception as e:
+- self.log_error('Error running sosreport: %s' % e)
++ self.log_error('Error running sos report: %s' % e)
+ raise
+
+ def retrieve_file(self, path):
+ """Copies the specified file from the host to our temp dir"""
+ destdir = self.tmpdir + '/'
+- dest = destdir + path.split('/')[-1]
++ dest = os.path.join(destdir, path.split('/')[-1])
+ try:
+- if not self.local:
+- if self.file_exists(path):
+- self.log_info("Copying remote %s to local %s" %
+- (path, destdir))
+- cmd = "/usr/bin/scp -oControlPath=%s %s@%s:%s %s" % (
+- self.control_path,
+- self.opts.ssh_user,
+- self.address,
+- path,
+- destdir
+- )
+- res = self.run_command(cmd, force_local=True)
+- return res['status'] == 0
+- else:
+- self.log_debug("Attempting to copy remote file %s, but it "
+- "does not exist on filesystem" % path)
+- return False
++ if self.file_exists(path):
++ self.log_info("Copying remote %s to local %s" %
++ (path, destdir))
++ self._transport.retrieve_file(path, dest)
+ else:
+- self.log_debug("Moving %s to %s" % (path, destdir))
+- shutil.copy(path, dest)
++ self.log_debug("Attempting to copy remote file %s, but it "
++ "does not exist on filesystem" % path)
++ return False
+ return True
+ except Exception as err:
+ self.log_debug("Failed to retrieve %s: %s" % (path, err))
+@@ -933,7 +745,7 @@ class SosNode():
+ """
+ path = ''.join(path.split())
+ try:
+- if len(path) <= 2: # ensure we have a non '/' path
++ if len(path.split('/')) <= 2: # ensure we have a non '/' path
+ self.log_debug("Refusing to remove path %s: appears to be "
+ "incorrect and possibly dangerous" % path)
+ return False
+@@ -959,14 +771,14 @@ class SosNode():
+ except Exception:
+ self.log_error('Failed to make archive readable')
+ return False
+- self.soslog.info('Retrieving sosreport from %s' % self.address)
+- self.ui_msg('Retrieving sosreport...')
++ self.soslog.info('Retrieving sos report from %s' % self.address)
++ self.ui_msg('Retrieving sos report...')
+ ret = self.retrieve_file(self.sos_path)
+ if ret:
+- self.ui_msg('Successfully collected sosreport')
++ self.ui_msg('Successfully collected sos report')
+ self.file_list.append(self.sos_path.split('/')[-1])
+ else:
+- self.log_error('Failed to retrieve sosreport')
++ self.log_error('Failed to retrieve sos report')
+ raise SystemExit
+ return True
+ else:
+@@ -976,8 +788,8 @@ class SosNode():
+ else:
+ e = [x.strip() for x in self.stdout.readlines() if x.strip][-1]
+ self.soslog.error(
+- 'Failed to run sosreport on %s: %s' % (self.address, e))
+- self.log_error('Failed to run sosreport. %s' % e)
++ 'Failed to run sos report on %s: %s' % (self.address, e))
++ self.log_error('Failed to run sos report. %s' % e)
+ return False
+
+ def remove_sos_archive(self):
+@@ -986,20 +798,20 @@ class SosNode():
+ if self.sos_path is None:
+ return
+ if 'sosreport' not in self.sos_path:
+- self.log_debug("Node sosreport path %s looks incorrect. Not "
++ self.log_debug("Node sos report path %s looks incorrect. Not "
+ "attempting to remove path" % self.sos_path)
+ return
+ removed = self.remove_file(self.sos_path)
+ if not removed:
+- self.log_error('Failed to remove sosreport')
++ self.log_error('Failed to remove sos report')
+
+ def cleanup(self):
+ """Remove the sos archive from the node once we have it locally"""
+ self.remove_sos_archive()
+ if self.sos_path:
+ for ext in ['.sha256', '.md5']:
+- if os.path.isfile(self.sos_path + ext):
+- self.remove_file(self.sos_path + ext)
++ if self.remove_file(self.sos_path + ext):
++ break
+ cleanup = self.host.set_cleanup_cmd()
+ if cleanup:
+ self.run_command(cleanup, need_root=True)
+@@ -1040,3 +852,5 @@ class SosNode():
+ msg = "Exception while making %s readable. Return code was %s"
+ self.log_error(msg % (filepath, res['status']))
+ raise Exception
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/transports/__init__.py b/sos/collector/transports/__init__.py
+new file mode 100644
+index 00000000..5be7dc6d
+--- /dev/null
++++ b/sos/collector/transports/__init__.py
+@@ -0,0 +1,317 @@
++# Copyright Red Hat 2021, Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++import inspect
++import logging
++import pexpect
++import re
++
++from pipes import quote
++from sos.collector.exceptions import (ConnectionException,
++ CommandTimeoutException)
++
++
++class RemoteTransport():
++ """The base class used for defining supported remote transports to connect
++ to remote nodes in conjunction with `sos collect`.
++
++ This abstraction is used to manage the backend connections to nodes so that
++ SoSNode() objects can be leveraged generically to connect to nodes, inspect
++ those nodes, and run commands on them.
++ """
++
++ name = 'undefined'
++
++ def __init__(self, address, commons):
++ self.address = address
++ self.opts = commons['cmdlineopts']
++ self.tmpdir = commons['tmpdir']
++ self.need_sudo = commons['need_sudo']
++ self._hostname = None
++ self.soslog = logging.getLogger('sos')
++ self.ui_log = logging.getLogger('sos_ui')
++
++ def _sanitize_log_msg(self, msg):
++ """Attempts to obfuscate sensitive information in log messages such as
++ passwords"""
++ reg = r'(?P<var>(pass|key|secret|PASS|KEY|SECRET).*?=)(?P<value>.*?\s)'
++ return re.sub(reg, r'\g<var>****** ', msg)
++
++ def log_info(self, msg):
++ """Used to print and log info messages"""
++ caller = inspect.stack()[1][3]
++ lmsg = '[%s:%s] %s' % (self.hostname, caller, msg)
++ self.soslog.info(lmsg)
++
++ def log_error(self, msg):
++ """Used to print and log error messages"""
++ caller = inspect.stack()[1][3]
++ lmsg = '[%s:%s] %s' % (self.hostname, caller, msg)
++ self.soslog.error(lmsg)
++
++ def log_debug(self, msg):
++ """Used to print and log debug messages"""
++ msg = self._sanitize_log_msg(msg)
++ caller = inspect.stack()[1][3]
++ msg = '[%s:%s] %s' % (self.hostname, caller, msg)
++ self.soslog.debug(msg)
++
++ @property
++ def hostname(self):
++ if self._hostname and 'localhost' not in self._hostname:
++ return self._hostname
++ return self.address
++
++ @property
++ def connected(self):
++ """Is the transport __currently__ connected to the node, or otherwise
++ capable of seamlessly running a command or similar on the node?
++ """
++ return False
++
++ @property
++ def remote_exec(self):
++ """This is the command string needed to leverage the remote transport
++ when executing commands. For example, for an SSH transport this would
++ be the `ssh <options>` string prepended to any command so that the
++ command is executed by the ssh binary.
++
++ This is also referenced by the `remote_exec` parameter for policies
++ when loading a policy for a remote node
++ """
++ return None
++
++ def connect(self, password):
++ """Perform the connection steps in order to ensure that we are able to
++ connect to the node for all future operations. Note that this should
++ not provide an interactive shell at this time.
++ """
++ if self._connect(password):
++ if not self._hostname:
++ self._get_hostname()
++ return True
++ return False
++
++ def _connect(self, password):
++ """Actually perform the connection requirements. Should be overridden
++ by specific transports that subclass RemoteTransport
++ """
++ raise NotImplementedError("Transport %s does not define connect"
++ % self.name)
++
++ def reconnect(self, password):
++ """Attempts to reconnect to the node using the standard connect()
++ but does not do so indefinitely. This imposes a strict number of retry
++ attempts before failing out
++ """
++ attempts = 1
++ last_err = 'unknown'
++ while attempts < 5:
++ self.log_debug("Attempting reconnect (#%s) to node" % attempts)
++ try:
++ if self.connect(password):
++ return True
++ except Exception as err:
++ self.log_debug("Attempt #%s exception: %s" % (attempts, err))
++ last_err = err
++ attempts += 1
++ self.log_error("Unable to reconnect to node after 5 attempts, "
++ "aborting.")
++ raise ConnectionException("last exception from transport: %s"
++ % last_err)
++
++ def disconnect(self):
++ """Perform whatever steps are necessary, if any, to terminate any
++ connection to the node
++ """
++ try:
++ if self._disconnect():
++ self.log_debug("Successfully disconnected from node")
++ else:
++ self.log_error("Unable to successfully disconnect, see log for"
++ " more details")
++ except Exception as err:
++ self.log_error("Failed to disconnect: %s" % err)
++
++ def _disconnect(self):
++ raise NotImplementedError("Transport %s does not define disconnect"
++ % self.name)
++
++ def run_command(self, cmd, timeout=180, need_root=False, env=None):
++ """Run a command on the node, returning its output and exit code.
++ This should return the exit code of the command being executed, not the
++ exit code of whatever mechanism the transport uses to execute that
++ command
++
++ :param cmd: The command to run
++ :type cmd: ``str``
++
++ :param timeout: The maximum time in seconds to allow the cmd to run
++ :type timeout: ``int``
++
++ :param get_pty: Does ``cmd`` require a pty?
++ :type get_pty: ``bool``
++
++ :param need_root: Does ``cmd`` require root privileges?
++ :type neeed_root: ``bool``
++
++ :param env: Specify env vars to be passed to the ``cmd``
++ :type env: ``dict``
++
++ :returns: Output of ``cmd`` and the exit code
++ :rtype: ``dict`` with keys ``output`` and ``status``
++ """
++ self.log_debug('Running command %s' % cmd)
++ # currently we only use/support the use of pexpect for handling the
++ # execution of these commands, as opposed to directly invoking
++ # subprocess.Popen() in conjunction with tools like sshpass.
++ # If that changes in the future, we'll add decision making logic here
++ # to route to the appropriate handler, but for now we just go straight
++ # to using pexpect
++ return self._run_command_with_pexpect(cmd, timeout, need_root, env)
++
++ def _format_cmd_for_exec(self, cmd):
++ """Format the command in the way needed for the remote transport to
++ successfully execute it as one would when manually executing it
++
++ :param cmd: The command being executed, as formatted by SoSNode
++ :type cmd: ``str``
++
++
++ :returns: The command further formatted as needed by this
++ transport
++ :rtype: ``str``
++ """
++ cmd = "%s %s" % (self.remote_exec, quote(cmd))
++ cmd = cmd.lstrip()
++ return cmd
++
++ def _run_command_with_pexpect(self, cmd, timeout, need_root, env):
++ """Execute the command using pexpect, which allows us to more easily
++ handle prompts and timeouts compared to directly leveraging the
++ subprocess.Popen() method.
++
++ :param cmd: The command to execute. This will be automatically
++ formatted to use the transport.
++ :type cmd: ``str``
++
++ :param timeout: The maximum time in seconds to run ``cmd``
++ :type timeout: ``int``
++
++ :param need_root: Does ``cmd`` need to run as root or with sudo?
++ :type need_root: ``bool``
++
++ :param env: Any env vars that ``cmd`` should be run with
++ :type env: ``dict``
++ """
++ cmd = self._format_cmd_for_exec(cmd)
++ result = pexpect.spawn(cmd, encoding='utf-8', env=env)
++
++ _expects = [pexpect.EOF, pexpect.TIMEOUT]
++ if need_root and self.opts.ssh_user != 'root':
++ _expects.extend([
++ '\\[sudo\\] password for .*:',
++ 'Password:'
++ ])
++
++ index = result.expect(_expects, timeout=timeout)
++
++ if index in [2, 3]:
++ self._send_pexpect_password(index, result)
++ index = result.expect(_expects, timeout=timeout)
++
++ if index == 0:
++ out = result.before
++ result.close()
++ return {'status': result.exitstatus, 'output': out}
++ elif index == 1:
++ raise CommandTimeoutException(cmd)
++
++ def _send_pexpect_password(self, index, result):
++ """Handle password prompts for sudo and su usage for non-root SSH users
++
++ :param index: The index pexpect.spawn returned to match against
++ either a sudo or su prompt
++ :type index: ``int``
++
++ :param result: The spawn running the command
++ :type result: ``pexpect.spawn``
++ """
++ if index == 2:
++ if not self.opts.sudo_pw and not self.opt.nopasswd_sudo:
++ msg = ("Unable to run command: sudo password "
++ "required but not provided")
++ self.log_error(msg)
++ raise Exception(msg)
++ result.sendline(self.opts.sudo_pw)
++ elif index == 3:
++ if not self.opts.root_password:
++ msg = ("Unable to run command as root: no root password given")
++ self.log_error(msg)
++ raise Exception(msg)
++ result.sendline(self.opts.root_password)
++
++ def _get_hostname(self):
++ """Determine the hostname of the node and set that for future reference
++ and logging
++
++ :returns: The hostname of the system, per the `hostname` command
++ :rtype: ``str``
++ """
++ _out = self.run_command('hostname')
++ if _out['status'] == 0:
++ self._hostname = _out['output'].strip()
++ self.log_info("Hostname set to %s" % self._hostname)
++ return self._hostname
++
++ def retrieve_file(self, fname, dest):
++ """Copy a remote file, fname, to dest on the local node
++
++ :param fname: The name of the file to retrieve
++ :type fname: ``str``
++
++ :param dest: Where to save the file to locally
++ :type dest: ``str``
++
++ :returns: True if file was successfully copied from remote, or False
++ :rtype: ``bool``
++ """
++ return self._retrieve_file(fname, dest)
++
++ def _retrieve_file(self, fname, dest):
++ raise NotImplementedError("Transport %s does not support file copying"
++ % self.name)
++
++ def read_file(self, fname):
++ """Read the given file fname and return its contents
++
++ :param fname: The name of the file to read
++ :type fname: ``str``
++
++ :returns: The content of the file
++ :rtype: ``str``
++ """
++ self.log_debug("Reading file %s" % fname)
++ return self._read_file(fname)
++
++ def _read_file(self, fname):
++ res = self.run_command("cat %s" % fname, timeout=5)
++ if res['status'] == 0:
++ return res['output']
++ else:
++ if 'No such file' in res['output']:
++ self.log_debug("File %s does not exist on node"
++ % fname)
++ else:
++ self.log_error("Error reading %s: %s" %
++ (fname, res['output'].split(':')[1:]))
++ return ''
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/transports/control_persist.py b/sos/collector/transports/control_persist.py
+new file mode 100644
+index 00000000..3e848b41
+--- /dev/null
++++ b/sos/collector/transports/control_persist.py
+@@ -0,0 +1,199 @@
++# Copyright Red Hat 2021, Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++
++import os
++import pexpect
++import subprocess
++
++from sos.collector.transports import RemoteTransport
++from sos.collector.exceptions import (InvalidPasswordException,
++ TimeoutPasswordAuthException,
++ PasswordRequestException,
++ AuthPermissionDeniedException,
++ ConnectionException,
++ ConnectionTimeoutException,
++ ControlSocketMissingException,
++ ControlPersistUnsupportedException)
++from sos.utilities import sos_get_command_output
++
++
++class SSHControlPersist(RemoteTransport):
++ """A transport for collect that leverages OpenSSH's Control Persist
++ functionality which uses control sockets to transparently keep a connection
++ open to the remote host without needing to rebuild the SSH connection for
++ each and every command executed on the node
++ """
++
++ name = 'control_persist'
++
++ def _check_for_control_persist(self):
++ """Checks to see if the local system supported SSH ControlPersist.
++
++ ControlPersist allows OpenSSH to keep a single open connection to a
++ remote host rather than building a new session each time. This is the
++ same feature that Ansible uses in place of paramiko, which we have a
++ need to drop in sos-collector.
++
++ This check relies on feedback from the ssh binary. The command being
++ run should always generate stderr output, but depending on what that
++ output reads we can determine if ControlPersist is supported or not.
++
++ For our purposes, a host that does not support ControlPersist is not
++ able to run sos-collector.
++
++ Returns
++ True if ControlPersist is supported, else raise Exception.
++ """
++ ssh_cmd = ['ssh', '-o', 'ControlPersist']
++ cmd = subprocess.Popen(ssh_cmd, stdout=subprocess.PIPE,
++ stderr=subprocess.PIPE)
++ out, err = cmd.communicate()
++ err = err.decode('utf-8')
++ if 'Bad configuration option' in err or 'Usage:' in err:
++ raise ControlPersistUnsupportedException
++ return True
++
++ def _connect(self, password=''):
++ """
++ Using ControlPersist, create the initial connection to the node.
++
++ This will generate an OpenSSH ControlPersist socket within the tmp
++ directory created or specified for sos-collector to use.
++
++ At most, we will wait 30 seconds for a connection. This involves a 15
++ second wait for the initial connection attempt, and a subsequent 15
++ second wait for a response when we supply a password.
++
++ Since we connect to nodes in parallel (using the --threads value), this
++ means that the time between 'Connecting to nodes...' and 'Beginning
++ collection of sosreports' that users see can be up to an amount of time
++ equal to 30*(num_nodes/threads) seconds.
++
++ Returns
++ True if session is successfully opened, else raise Exception
++ """
++ try:
++ self._check_for_control_persist()
++ except ControlPersistUnsupportedException:
++ self.log_error("OpenSSH ControlPersist is not locally supported. "
++ "Please update your OpenSSH installation.")
++ raise
++ self.log_info('Opening SSH session to create control socket')
++ self.control_path = ("%s/.sos-collector-%s" % (self.tmpdir,
++ self.address))
++ self.ssh_cmd = ''
++ connected = False
++ ssh_key = ''
++ ssh_port = ''
++ if self.opts.ssh_port != 22:
++ ssh_port = "-p%s " % self.opts.ssh_port
++ if self.opts.ssh_key:
++ ssh_key = "-i%s" % self.opts.ssh_key
++
++ cmd = ("ssh %s %s -oControlPersist=600 -oControlMaster=auto "
++ "-oStrictHostKeyChecking=no -oControlPath=%s %s@%s "
++ "\"echo Connected\"" % (ssh_key,
++ ssh_port,
++ self.control_path,
++ self.opts.ssh_user,
++ self.address))
++ res = pexpect.spawn(cmd, encoding='utf-8')
++
++ connect_expects = [
++ u'Connected',
++ u'password:',
++ u'.*Permission denied.*',
++ u'.* port .*: No route to host',
++ u'.*Could not resolve hostname.*',
++ pexpect.TIMEOUT
++ ]
++
++ index = res.expect(connect_expects, timeout=15)
++
++ if index == 0:
++ connected = True
++ elif index == 1:
++ if password:
++ pass_expects = [
++ u'Connected',
++ u'Permission denied, please try again.',
++ pexpect.TIMEOUT
++ ]
++ res.sendline(password)
++ pass_index = res.expect(pass_expects, timeout=15)
++ if pass_index == 0:
++ connected = True
++ elif pass_index == 1:
++ # Note that we do not get an exitstatus here, so matching
++ # this line means an invalid password will be reported for
++ # both invalid passwords and invalid user names
++ raise InvalidPasswordException
++ elif pass_index == 2:
++ raise TimeoutPasswordAuthException
++ else:
++ raise PasswordRequestException
++ elif index == 2:
++ raise AuthPermissionDeniedException
++ elif index == 3:
++ raise ConnectionException(self.address, self.opts.ssh_port)
++ elif index == 4:
++ raise ConnectionException(self.address)
++ elif index == 5:
++ raise ConnectionTimeoutException
++ else:
++ raise Exception("Unknown error, client returned %s" % res.before)
++ if connected:
++ if not os.path.exists(self.control_path):
++ raise ControlSocketMissingException
++ self.log_debug("Successfully created control socket at %s"
++ % self.control_path)
++ return True
++ return False
++
++ def _disconnect(self):
++ if os.path.exists(self.control_path):
++ try:
++ os.remove(self.control_path)
++ return True
++ except Exception as err:
++ self.log_debug("Could not disconnect properly: %s" % err)
++ return False
++ self.log_debug("Control socket not present when attempting to "
++ "terminate session")
++
++ @property
++ def connected(self):
++ """Check if the SSH control socket exists
++
++ The control socket is automatically removed by the SSH daemon in the
++ event that the last connection to the node was greater than the timeout
++ set by the ControlPersist option. This can happen for us if we are
++ collecting from a large number of nodes, and the timeout expires before
++ we start collection.
++ """
++ return os.path.exists(self.control_path)
++
++ @property
++ def remote_exec(self):
++ if not self.ssh_cmd:
++ self.ssh_cmd = "ssh -oControlPath=%s %s@%s" % (
++ self.control_path, self.opts.ssh_user, self.address
++ )
++ return self.ssh_cmd
++
++ def _retrieve_file(self, fname, dest):
++ cmd = "/usr/bin/scp -oControlPath=%s %s@%s:%s %s" % (
++ self.control_path, self.opts.ssh_user, self.address, fname, dest
++ )
++ res = sos_get_command_output(cmd)
++ return res['status'] == 0
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/transports/local.py b/sos/collector/transports/local.py
+new file mode 100644
+index 00000000..a4897f19
+--- /dev/null
++++ b/sos/collector/transports/local.py
+@@ -0,0 +1,49 @@
++# Copyright Red Hat 2021, Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++import os
++import shutil
++
++from sos.collector.transports import RemoteTransport
++
++
++class LocalTransport(RemoteTransport):
++ """A 'transport' to represent a local node. This allows us to more easily
++ extend SoSNode() without having a ton of 'if local' or similar checks in
++ more places than we actually need them
++ """
++
++ name = 'local_node'
++
++ def _connect(self, password):
++ return True
++
++ def _disconnect(self):
++ return True
++
++ @property
++ def connected(self):
++ return True
++
++ def _retrieve_file(self, fname, dest):
++ self.log_debug("Moving %s to %s" % (fname, dest))
++ shutil.copy(fname, dest)
++
++ def _format_cmd_for_exec(self, cmd):
++ return cmd
++
++ def _read_file(self, fname):
++ if os.path.exists(fname):
++ with open(fname, 'r') as rfile:
++ return rfile.read()
++ self.log_debug("No such file: %s" % fname)
++ return ''
++
++# vim: set et ts=4 sw=4 :
+--
+2.31.1
+
+From 07d96d52ef69b9f8fe1ef32a1b88089d31c33fe8 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 27 Sep 2021 12:28:27 -0400
+Subject: [PATCH 2/2] [plugins] Update plugins to use new os.path.join wrapper
+
+Updates plugins to use the new `self.path_join()` wrapper for
+`os.path.join()` so that these plugins now account for non-/ sysroots
+for their collections.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/__init__.py | 2 +-
+ sos/report/plugins/azure.py | 4 +--
+ sos/report/plugins/collectd.py | 2 +-
+ sos/report/plugins/container_log.py | 2 +-
+ sos/report/plugins/corosync.py | 2 +-
+ sos/report/plugins/docker_distribution.py | 5 ++--
+ sos/report/plugins/ds.py | 3 +--
+ sos/report/plugins/elastic.py | 4 ++-
+ sos/report/plugins/etcd.py | 2 +-
+ sos/report/plugins/gluster.py | 3 ++-
+ sos/report/plugins/jars.py | 2 +-
+ sos/report/plugins/kdump.py | 4 +--
+ sos/report/plugins/libvirt.py | 2 +-
+ sos/report/plugins/logs.py | 8 +++---
+ sos/report/plugins/manageiq.py | 12 ++++-----
+ sos/report/plugins/numa.py | 9 +++----
+ sos/report/plugins/openstack_instack.py | 2 +-
+ sos/report/plugins/openstack_nova.py | 2 +-
+ sos/report/plugins/openvswitch.py | 13 ++++-----
+ sos/report/plugins/origin.py | 28 +++++++++++---------
+ sos/report/plugins/ovirt.py | 2 +-
+ sos/report/plugins/ovirt_engine_backup.py | 5 ++--
+ sos/report/plugins/ovn_central.py | 26 +++++++++---------
+ sos/report/plugins/ovn_host.py | 4 +--
+ sos/report/plugins/pacemaker.py | 4 +--
+ sos/report/plugins/pcp.py | 32 +++++++++++------------
+ sos/report/plugins/postfix.py | 2 +-
+ sos/report/plugins/postgresql.py | 2 +-
+ sos/report/plugins/powerpc.py | 2 +-
+ sos/report/plugins/processor.py | 3 +--
+ sos/report/plugins/python.py | 4 +--
+ sos/report/plugins/sar.py | 5 ++--
+ sos/report/plugins/sos_extras.py | 2 +-
+ sos/report/plugins/ssh.py | 7 +++--
+ sos/report/plugins/unpackaged.py | 4 +--
+ sos/report/plugins/watchdog.py | 13 +++++----
+ sos/report/plugins/yum.py | 2 +-
+ 37 files changed, 115 insertions(+), 115 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 1f84bca4..ec138f83 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -2897,7 +2897,7 @@ class Plugin():
+ try:
+ cmd_line_paths = glob.glob(cmd_line_glob)
+ for path in cmd_line_paths:
+- f = open(path, 'r')
++ f = open(self.path_join(path), 'r')
+ cmd_line = f.read().strip()
+ if process in cmd_line:
+ status = True
+diff --git a/sos/report/plugins/azure.py b/sos/report/plugins/azure.py
+index 45971a61..90999b3f 100644
+--- a/sos/report/plugins/azure.py
++++ b/sos/report/plugins/azure.py
+@@ -8,8 +8,8 @@
+ #
+ # See the LICENSE file in the source distribution for further information.
+
+-import os
+ from sos.report.plugins import Plugin, UbuntuPlugin, RedHatPlugin
++import os
+
+
+ class Azure(Plugin, UbuntuPlugin):
+@@ -38,7 +38,7 @@ class Azure(Plugin, UbuntuPlugin):
+
+ for path, subdirs, files in os.walk("/var/log/azure"):
+ for name in files:
+- self.add_copy_spec(os.path.join(path, name), sizelimit=limit)
++ self.add_copy_spec(self.path_join(path, name), sizelimit=limit)
+
+ self.add_cmd_output((
+ 'curl -s -H Metadata:true '
+diff --git a/sos/report/plugins/collectd.py b/sos/report/plugins/collectd.py
+index 80d4b00a..8584adf9 100644
+--- a/sos/report/plugins/collectd.py
++++ b/sos/report/plugins/collectd.py
+@@ -33,7 +33,7 @@ class Collectd(Plugin, IndependentPlugin):
+
+ p = re.compile('^LoadPlugin.*')
+ try:
+- with open("/etc/collectd.conf") as f:
++ with open(self.path_join("/etc/collectd.conf"), 'r') as f:
+ for line in f:
+ if p.match(line):
+ self.add_alert("Active Plugin found: %s" %
+diff --git a/sos/report/plugins/container_log.py b/sos/report/plugins/container_log.py
+index 14e0b7d8..e8dedad2 100644
+--- a/sos/report/plugins/container_log.py
++++ b/sos/report/plugins/container_log.py
+@@ -29,6 +29,6 @@ class ContainerLog(Plugin, IndependentPlugin):
+ """Collect *.log files from subdirs of passed root path
+ """
+ for dirName, _, _ in os.walk(root):
+- self.add_copy_spec(os.path.join(dirName, '*.log'))
++ self.add_copy_spec(self.path_join(dirName, '*.log'))
+
+ # vim: set et ts=4 sw=4 :
+diff --git a/sos/report/plugins/corosync.py b/sos/report/plugins/corosync.py
+index d74086e3..10e096c6 100644
+--- a/sos/report/plugins/corosync.py
++++ b/sos/report/plugins/corosync.py
+@@ -47,7 +47,7 @@ class Corosync(Plugin):
+ # (it isnt precise but sufficient)
+ pattern = r'^\s*(logging.)?logfile:\s*(\S+)$'
+ try:
+- with open("/etc/corosync/corosync.conf") as f:
++ with open(self.path_join("/etc/corosync/corosync.conf"), 'r') as f:
+ for line in f:
+ if re.match(pattern, line):
+ self.add_copy_spec(re.search(pattern, line).group(2))
+diff --git a/sos/report/plugins/docker_distribution.py b/sos/report/plugins/docker_distribution.py
+index 84222ff7..e760f252 100644
+--- a/sos/report/plugins/docker_distribution.py
++++ b/sos/report/plugins/docker_distribution.py
+@@ -19,8 +19,9 @@ class DockerDistribution(Plugin):
+ def setup(self):
+ self.add_copy_spec('/etc/docker-distribution/')
+ self.add_journal('docker-distribution')
+- if self.path_exists('/etc/docker-distribution/registry/config.yml'):
+- with open('/etc/docker-distribution/registry/config.yml') as f:
++ conf = self.path_join('/etc/docker-distribution/registry/config.yml')
++ if self.path_exists(conf):
++ with open(conf) as f:
+ for line in f:
+ if 'rootdirectory' in line:
+ loc = line.split()[1]
+diff --git a/sos/report/plugins/ds.py b/sos/report/plugins/ds.py
+index addf49e1..43feb21e 100644
+--- a/sos/report/plugins/ds.py
++++ b/sos/report/plugins/ds.py
+@@ -11,7 +11,6 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ from sos.report.plugins import Plugin, RedHatPlugin
+-import os
+
+
+ class DirectoryServer(Plugin, RedHatPlugin):
+@@ -47,7 +46,7 @@ class DirectoryServer(Plugin, RedHatPlugin):
+ try:
+ for d in self.listdir("/etc/dirsrv"):
+ if d[0:5] == 'slapd':
+- certpath = os.path.join("/etc/dirsrv", d)
++ certpath = self.path_join("/etc/dirsrv", d)
+ self.add_cmd_output("certutil -L -d %s" % certpath)
+ self.add_cmd_output("dsctl %s healthcheck" % d)
+ except OSError:
+diff --git a/sos/report/plugins/elastic.py b/sos/report/plugins/elastic.py
+index ad9a06ff..da2662bc 100644
+--- a/sos/report/plugins/elastic.py
++++ b/sos/report/plugins/elastic.py
+@@ -39,7 +39,9 @@ class Elastic(Plugin, IndependentPlugin):
+ return hostname, port
+
+ def setup(self):
+- els_config_file = "/etc/elasticsearch/elasticsearch.yml"
++ els_config_file = self.path_join(
++ "/etc/elasticsearch/elasticsearch.yml"
++ )
+ self.add_copy_spec(els_config_file)
+
+ if self.get_option("all_logs"):
+diff --git a/sos/report/plugins/etcd.py b/sos/report/plugins/etcd.py
+index fd4f67eb..fe017e9f 100644
+--- a/sos/report/plugins/etcd.py
++++ b/sos/report/plugins/etcd.py
+@@ -62,7 +62,7 @@ class etcd(Plugin, RedHatPlugin):
+
+ def get_etcd_url(self):
+ try:
+- with open('/etc/etcd/etcd.conf', 'r') as ef:
++ with open(self.path_join('/etc/etcd/etcd.conf'), 'r') as ef:
+ for line in ef:
+ if line.startswith('ETCD_LISTEN_CLIENT_URLS'):
+ return line.split('=')[1].replace('"', '').strip()
+diff --git a/sos/report/plugins/gluster.py b/sos/report/plugins/gluster.py
+index a44ffeb7..e518e3d3 100644
+--- a/sos/report/plugins/gluster.py
++++ b/sos/report/plugins/gluster.py
+@@ -35,9 +35,10 @@ class Gluster(Plugin, RedHatPlugin):
+ ]
+ for statedump_file in statedump_entries:
+ statedumps_present = statedumps_present+1
++ _spath = self.path_join(name_dir, statedump_file)
+ ret = -1
+ while ret == -1:
+- with open(name_dir + '/' + statedump_file, 'r') as sfile:
++ with open(_spath, 'r') as sfile:
+ last_line = sfile.readlines()[-1]
+ ret = string.count(last_line, 'DUMP_END_TIME')
+
+diff --git a/sos/report/plugins/jars.py b/sos/report/plugins/jars.py
+index 0d3cf37e..4b98684e 100644
+--- a/sos/report/plugins/jars.py
++++ b/sos/report/plugins/jars.py
+@@ -63,7 +63,7 @@ class Jars(Plugin, RedHatPlugin):
+ for location in locations:
+ for dirpath, _, filenames in os.walk(location):
+ for filename in filenames:
+- path = os.path.join(dirpath, filename)
++ path = self.path_join(dirpath, filename)
+ if Jars.is_jar(path):
+ jar_paths.append(path)
+
+diff --git a/sos/report/plugins/kdump.py b/sos/report/plugins/kdump.py
+index 757c2736..66565664 100644
+--- a/sos/report/plugins/kdump.py
++++ b/sos/report/plugins/kdump.py
+@@ -40,7 +40,7 @@ class RedHatKDump(KDump, RedHatPlugin):
+ packages = ('kexec-tools',)
+
+ def fstab_parse_fs(self, device):
+- with open('/etc/fstab', 'r') as fp:
++ with open(self.path_join('/etc/fstab'), 'r') as fp:
+ for line in fp:
+ if line.startswith((device)):
+ return line.split()[1].rstrip('/')
+@@ -50,7 +50,7 @@ class RedHatKDump(KDump, RedHatPlugin):
+ fs = ""
+ path = "/var/crash"
+
+- with open('/etc/kdump.conf', 'r') as fp:
++ with open(self.path_join('/etc/kdump.conf'), 'r') as fp:
+ for line in fp:
+ if line.startswith("path"):
+ path = line.split()[1]
+diff --git a/sos/report/plugins/libvirt.py b/sos/report/plugins/libvirt.py
+index be8120ff..5caa5802 100644
+--- a/sos/report/plugins/libvirt.py
++++ b/sos/report/plugins/libvirt.py
+@@ -55,7 +55,7 @@ class Libvirt(Plugin, IndependentPlugin):
+ else:
+ self.add_copy_spec("/var/log/libvirt")
+
+- if self.path_exists(self.join_sysroot(libvirt_keytab)):
++ if self.path_exists(self.path_join(libvirt_keytab)):
+ self.add_cmd_output("klist -ket %s" % libvirt_keytab)
+
+ self.add_cmd_output("ls -lR /var/lib/libvirt/qemu")
+diff --git a/sos/report/plugins/logs.py b/sos/report/plugins/logs.py
+index ee6bb98d..606e574a 100644
+--- a/sos/report/plugins/logs.py
++++ b/sos/report/plugins/logs.py
+@@ -24,15 +24,15 @@ class Logs(Plugin, IndependentPlugin):
+ since = self.get_option("since")
+
+ if self.path_exists('/etc/rsyslog.conf'):
+- with open('/etc/rsyslog.conf', 'r') as conf:
++ with open(self.path_join('/etc/rsyslog.conf'), 'r') as conf:
+ for line in conf.readlines():
+ if line.startswith('$IncludeConfig'):
+ confs += glob.glob(line.split()[1])
+
+ for conf in confs:
+- if not self.path_exists(conf):
++ if not self.path_exists(self.path_join(conf)):
+ continue
+- config = self.join_sysroot(conf)
++ config = self.path_join(conf)
+ logs += self.do_regex_find_all(r"^\S+\s+(-?\/.*$)\s+", config)
+
+ for i in logs:
+@@ -60,7 +60,7 @@ class Logs(Plugin, IndependentPlugin):
+ # - there is some data present, either persistent or runtime only
+ # - systemd-journald service exists
+ # otherwise fallback to collecting few well known logfiles directly
+- journal = any([self.path_exists(p + "/log/journal/")
++ journal = any([self.path_exists(self.path_join(p, "log/journal/"))
+ for p in ["/var", "/run"]])
+ if journal and self.is_service("systemd-journald"):
+ self.add_journal(since=since, tags='journal_full', priority=100)
+diff --git a/sos/report/plugins/manageiq.py b/sos/report/plugins/manageiq.py
+index 27ad6ef4..e20c4a2a 100644
+--- a/sos/report/plugins/manageiq.py
++++ b/sos/report/plugins/manageiq.py
+@@ -58,7 +58,7 @@ class ManageIQ(Plugin, RedHatPlugin):
+ # Log files to collect from miq_dir/log/
+ miq_log_dir = os.path.join(miq_dir, "log")
+
+- miq_main_log_files = [
++ miq_main_logs = [
+ 'ansible_tower.log',
+ 'top_output.log',
+ 'evm.log',
+@@ -81,16 +81,16 @@ class ManageIQ(Plugin, RedHatPlugin):
+ self.add_copy_spec(list(self.files))
+
+ self.add_copy_spec([
+- os.path.join(self.miq_conf_dir, x) for x in self.miq_conf_files
++ self.path_join(self.miq_conf_dir, x) for x in self.miq_conf_files
+ ])
+
+ # Collect main log files without size limit.
+ self.add_copy_spec([
+- os.path.join(self.miq_log_dir, x) for x in self.miq_main_log_files
++ self.path_join(self.miq_log_dir, x) for x in self.miq_main_logs
+ ], sizelimit=0)
+
+ self.add_copy_spec([
+- os.path.join(self.miq_log_dir, x) for x in self.miq_log_files
++ self.path_join(self.miq_log_dir, x) for x in self.miq_log_files
+ ])
+
+ self.add_copy_spec([
+@@ -101,8 +101,8 @@ class ManageIQ(Plugin, RedHatPlugin):
+ if environ.get("APPLIANCE_PG_DATA"):
+ pg_dir = environ.get("APPLIANCE_PG_DATA")
+ self.add_copy_spec([
+- os.path.join(pg_dir, 'pg_log'),
+- os.path.join(pg_dir, 'postgresql.conf')
++ self.path_join(pg_dir, 'pg_log'),
++ self.path_join(pg_dir, 'postgresql.conf')
+ ])
+
+ # vim: set et ts=4 sw=4 :
+diff --git a/sos/report/plugins/numa.py b/sos/report/plugins/numa.py
+index 0faef8d2..9094baef 100644
+--- a/sos/report/plugins/numa.py
++++ b/sos/report/plugins/numa.py
+@@ -9,7 +9,6 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ from sos.report.plugins import Plugin, IndependentPlugin
+-import os.path
+
+
+ class Numa(Plugin, IndependentPlugin):
+@@ -42,10 +41,10 @@ class Numa(Plugin, IndependentPlugin):
+ ])
+
+ self.add_copy_spec([
+- os.path.join(numa_path, "node*/meminfo"),
+- os.path.join(numa_path, "node*/cpulist"),
+- os.path.join(numa_path, "node*/distance"),
+- os.path.join(numa_path, "node*/hugepages/hugepages-*/*")
++ self.path_join(numa_path, "node*/meminfo"),
++ self.path_join(numa_path, "node*/cpulist"),
++ self.path_join(numa_path, "node*/distance"),
++ self.path_join(numa_path, "node*/hugepages/hugepages-*/*")
+ ])
+
+ # vim: set et ts=4 sw=4 :
+diff --git a/sos/report/plugins/openstack_instack.py b/sos/report/plugins/openstack_instack.py
+index 7c56c162..5b4f7d41 100644
+--- a/sos/report/plugins/openstack_instack.py
++++ b/sos/report/plugins/openstack_instack.py
+@@ -68,7 +68,7 @@ class OpenStackInstack(Plugin):
+ p = uc_config.get(opt)
+ if p:
+ if not os.path.isabs(p):
+- p = os.path.join('/home/stack', p)
++ p = self.path_join('/home/stack', p)
+ self.add_copy_spec(p)
+ except Exception:
+ pass
+diff --git a/sos/report/plugins/openstack_nova.py b/sos/report/plugins/openstack_nova.py
+index 53210c48..f840081e 100644
+--- a/sos/report/plugins/openstack_nova.py
++++ b/sos/report/plugins/openstack_nova.py
+@@ -103,7 +103,7 @@ class OpenStackNova(Plugin):
+ "nova-scheduler.log*"
+ ]
+ for novalog in novalogs:
+- self.add_copy_spec(os.path.join(novadir, novalog))
++ self.add_copy_spec(self.path_join(novadir, novalog))
+
+ self.add_copy_spec([
+ "/etc/nova/",
+diff --git a/sos/report/plugins/openvswitch.py b/sos/report/plugins/openvswitch.py
+index 003596c6..179d1532 100644
+--- a/sos/report/plugins/openvswitch.py
++++ b/sos/report/plugins/openvswitch.py
+@@ -10,7 +10,6 @@
+
+ from sos.report.plugins import Plugin, RedHatPlugin, DebianPlugin, UbuntuPlugin
+
+-from os.path import join as path_join
+ from os import environ
+
+ import re
+@@ -65,7 +64,9 @@ class OpenVSwitch(Plugin):
+ log_dirs.append(environ.get('OVS_LOGDIR'))
+
+ if not all_logs:
+- self.add_copy_spec([path_join(ld, '*.log') for ld in log_dirs])
++ self.add_copy_spec([
++ self.path_join(ld, '*.log') for ld in log_dirs
++ ])
+ else:
+ self.add_copy_spec(log_dirs)
+
+@@ -76,13 +77,13 @@ class OpenVSwitch(Plugin):
+ ])
+
+ self.add_copy_spec([
+- path_join('/usr/local/etc/openvswitch', 'conf.db'),
+- path_join('/etc/openvswitch', 'conf.db'),
+- path_join('/var/lib/openvswitch', 'conf.db'),
++ self.path_join('/usr/local/etc/openvswitch', 'conf.db'),
++ self.path_join('/etc/openvswitch', 'conf.db'),
++ self.path_join('/var/lib/openvswitch', 'conf.db'),
+ ])
+ ovs_dbdir = environ.get('OVS_DBDIR')
+ if ovs_dbdir:
+- self.add_copy_spec(path_join(ovs_dbdir, 'conf.db'))
++ self.add_copy_spec(self.path_join(ovs_dbdir, 'conf.db'))
+
+ self.add_cmd_output([
+ # The '-t 5' adds an upper bound on how long to wait to connect
+diff --git a/sos/report/plugins/origin.py b/sos/report/plugins/origin.py
+index f9cc32c1..7df9c019 100644
+--- a/sos/report/plugins/origin.py
++++ b/sos/report/plugins/origin.py
+@@ -69,20 +69,21 @@ class OpenShiftOrigin(Plugin):
+
+ def is_static_etcd(self):
+ """Determine if we are on a node running etcd"""
+- return self.path_exists(os.path.join(self.static_pod_dir, "etcd.yaml"))
++ return self.path_exists(self.path_join(self.static_pod_dir,
++ "etcd.yaml"))
+
+ def is_static_pod_compatible(self):
+ """Determine if a node is running static pods"""
+ return self.path_exists(self.static_pod_dir)
+
+ def setup(self):
+- bstrap_node_cfg = os.path.join(self.node_base_dir,
+- "bootstrap-" + self.node_cfg_file)
+- bstrap_kubeconfig = os.path.join(self.node_base_dir,
+- "bootstrap.kubeconfig")
+- node_certs = os.path.join(self.node_base_dir, "certs", "*")
+- node_client_ca = os.path.join(self.node_base_dir, "client-ca.crt")
+- admin_cfg = os.path.join(self.master_base_dir, "admin.kubeconfig")
++ bstrap_node_cfg = self.path_join(self.node_base_dir,
++ "bootstrap-" + self.node_cfg_file)
++ bstrap_kubeconfig = self.path_join(self.node_base_dir,
++ "bootstrap.kubeconfig")
++ node_certs = self.path_join(self.node_base_dir, "certs", "*")
++ node_client_ca = self.path_join(self.node_base_dir, "client-ca.crt")
++ admin_cfg = self.path_join(self.master_base_dir, "admin.kubeconfig")
+ oc_cmd_admin = "%s --config=%s" % ("oc", admin_cfg)
+ static_pod_logs_cmd = "master-logs"
+
+@@ -92,11 +93,12 @@ class OpenShiftOrigin(Plugin):
+ self.add_copy_spec([
+ self.master_cfg,
+ self.master_env,
+- os.path.join(self.master_base_dir, "*.crt"),
++ self.path_join(self.master_base_dir, "*.crt"),
+ ])
+
+ if self.is_static_pod_compatible():
+- self.add_copy_spec(os.path.join(self.static_pod_dir, "*.yaml"))
++ self.add_copy_spec(self.path_join(self.static_pod_dir,
++ "*.yaml"))
+ self.add_cmd_output([
+ "%s api api" % static_pod_logs_cmd,
+ "%s controllers controllers" % static_pod_logs_cmd,
+@@ -177,9 +179,9 @@ class OpenShiftOrigin(Plugin):
+ node_client_ca,
+ bstrap_node_cfg,
+ bstrap_kubeconfig,
+- os.path.join(self.node_base_dir, "*.crt"),
+- os.path.join(self.node_base_dir, "resolv.conf"),
+- os.path.join(self.node_base_dir, "node-dnsmasq.conf"),
++ self.path_join(self.node_base_dir, "*.crt"),
++ self.path_join(self.node_base_dir, "resolv.conf"),
++ self.path_join(self.node_base_dir, "node-dnsmasq.conf"),
+ ])
+
+ self.add_journal(units="atomic-openshift-node")
+diff --git a/sos/report/plugins/ovirt.py b/sos/report/plugins/ovirt.py
+index 1de606be..09647bf1 100644
+--- a/sos/report/plugins/ovirt.py
++++ b/sos/report/plugins/ovirt.py
+@@ -216,7 +216,7 @@ class Ovirt(Plugin, RedHatPlugin):
+ "isouploader.conf"
+ ]
+ for conf_file in passwd_files:
+- conf_path = os.path.join("/etc/ovirt-engine", conf_file)
++ conf_path = self.path_join("/etc/ovirt-engine", conf_file)
+ self.do_file_sub(
+ conf_path,
+ r"passwd=(.*)",
+diff --git a/sos/report/plugins/ovirt_engine_backup.py b/sos/report/plugins/ovirt_engine_backup.py
+index 676e419e..7fb6a5c7 100644
+--- a/sos/report/plugins/ovirt_engine_backup.py
++++ b/sos/report/plugins/ovirt_engine_backup.py
+@@ -8,7 +8,6 @@
+ #
+ # See the LICENSE file in the source distribution for further information.
+
+-import os
+ from sos.report.plugins import (Plugin, RedHatPlugin)
+ from datetime import datetime
+
+@@ -29,11 +28,11 @@ class oVirtEngineBackup(Plugin, RedHatPlugin):
+
+ def setup(self):
+ now = datetime.now().strftime("%Y%m%d%H%M%S")
+- backup_filename = os.path.join(
++ backup_filename = self.path_join(
+ self.get_option("backupdir"),
+ "engine-db-backup-%s.tar.gz" % (now)
+ )
+- log_filename = os.path.join(
++ log_filename = self.path_join(
+ self.get_option("backupdir"),
+ "engine-db-backup-%s.log" % (now)
+ )
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index d6647aad..914eda60 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -42,7 +42,7 @@ class OVNCentral(Plugin):
+ return
+ else:
+ try:
+- with open(filename, 'r') as f:
++ with open(self.path_join(filename), 'r') as f:
+ try:
+ db = json.load(f)
+ except Exception:
+@@ -71,13 +71,13 @@ class OVNCentral(Plugin):
+ ovs_rundir = os.environ.get('OVS_RUNDIR')
+ for pidfile in ['ovnnb_db.pid', 'ovnsb_db.pid', 'ovn-northd.pid']:
+ self.add_copy_spec([
+- os.path.join('/var/lib/openvswitch/ovn', pidfile),
+- os.path.join('/usr/local/var/run/openvswitch', pidfile),
+- os.path.join('/run/openvswitch/', pidfile),
++ self.path_join('/var/lib/openvswitch/ovn', pidfile),
++ self.path_join('/usr/local/var/run/openvswitch', pidfile),
++ self.path_join('/run/openvswitch/', pidfile),
+ ])
+
+ if ovs_rundir:
+- self.add_copy_spec(os.path.join(ovs_rundir, pidfile))
++ self.add_copy_spec(self.path_join(ovs_rundir, pidfile))
+
+ if self.get_option("all_logs"):
+ self.add_copy_spec("/var/log/ovn/")
+@@ -104,7 +104,7 @@ class OVNCentral(Plugin):
+
+ schema_dir = '/usr/share/openvswitch'
+
+- nb_tables = self.get_tables_from_schema(os.path.join(
++ nb_tables = self.get_tables_from_schema(self.path_join(
+ schema_dir, 'ovn-nb.ovsschema'))
+
+ self.add_database_output(nb_tables, nbctl_cmds, 'ovn-nbctl')
+@@ -116,7 +116,7 @@ class OVNCentral(Plugin):
+ format(self.ovn_sbdb_sock_path),
+ "output": "Leader: self"}
+ if self.test_predicate(self, pred=SoSPredicate(self, cmd_outputs=co)):
+- sb_tables = self.get_tables_from_schema(os.path.join(
++ sb_tables = self.get_tables_from_schema(self.path_join(
+ schema_dir, 'ovn-sb.ovsschema'), ['Logical_Flow'])
+ self.add_database_output(sb_tables, sbctl_cmds, 'ovn-sbctl')
+ cmds += sbctl_cmds
+@@ -134,14 +134,14 @@ class OVNCentral(Plugin):
+ ovs_dbdir = os.environ.get('OVS_DBDIR')
+ for dbfile in ['ovnnb_db.db', 'ovnsb_db.db']:
+ self.add_copy_spec([
+- os.path.join('/var/lib/openvswitch/ovn', dbfile),
+- os.path.join('/usr/local/etc/openvswitch', dbfile),
+- os.path.join('/etc/openvswitch', dbfile),
+- os.path.join('/var/lib/openvswitch', dbfile),
+- os.path.join('/var/lib/ovn/etc', dbfile),
++ self.path_join('/var/lib/openvswitch/ovn', dbfile),
++ self.path_join('/usr/local/etc/openvswitch', dbfile),
++ self.path_join('/etc/openvswitch', dbfile),
++ self.path_join('/var/lib/openvswitch', dbfile),
++ self.path_join('/var/lib/ovn/etc', dbfile)
+ ])
+ if ovs_dbdir:
+- self.add_copy_spec(os.path.join(ovs_dbdir, dbfile))
++ self.add_copy_spec(self.path_join(ovs_dbdir, dbfile))
+
+ self.add_journal(units="ovn-northd")
+
+diff --git a/sos/report/plugins/ovn_host.py b/sos/report/plugins/ovn_host.py
+index 3742c49f..78604a15 100644
+--- a/sos/report/plugins/ovn_host.py
++++ b/sos/report/plugins/ovn_host.py
+@@ -35,7 +35,7 @@ class OVNHost(Plugin):
+ else:
+ self.add_copy_spec("/var/log/ovn/*.log")
+
+- self.add_copy_spec([os.path.join(pp, pidfile) for pp in pid_paths])
++ self.add_copy_spec([self.path_join(pp, pidfile) for pp in pid_paths])
+
+ self.add_copy_spec('/etc/sysconfig/ovn-controller')
+
+@@ -49,7 +49,7 @@ class OVNHost(Plugin):
+
+ def check_enabled(self):
+ return (any([self.path_isfile(
+- os.path.join(pp, pidfile)) for pp in pid_paths]) or
++ self.path_join(pp, pidfile)) for pp in pid_paths]) or
+ super(OVNHost, self).check_enabled())
+
+
+diff --git a/sos/report/plugins/pacemaker.py b/sos/report/plugins/pacemaker.py
+index 497807ff..6ce80881 100644
+--- a/sos/report/plugins/pacemaker.py
++++ b/sos/report/plugins/pacemaker.py
+@@ -129,7 +129,7 @@ class Pacemaker(Plugin):
+
+ class DebianPacemaker(Pacemaker, DebianPlugin, UbuntuPlugin):
+ def setup(self):
+- self.envfile = "/etc/default/pacemaker"
++ self.envfile = self.path_join("/etc/default/pacemaker")
+ self.setup_crm_shell()
+ self.setup_pcs()
+ super(DebianPacemaker, self).setup()
+@@ -141,7 +141,7 @@ class DebianPacemaker(Pacemaker, DebianPlugin, UbuntuPlugin):
+
+ class RedHatPacemaker(Pacemaker, RedHatPlugin):
+ def setup(self):
+- self.envfile = "/etc/sysconfig/pacemaker"
++ self.envfile = self.path_join("/etc/sysconfig/pacemaker")
+ self.setup_pcs()
+ self.add_copy_spec("/etc/sysconfig/sbd")
+ super(RedHatPacemaker, self).setup()
+diff --git a/sos/report/plugins/pcp.py b/sos/report/plugins/pcp.py
+index 9707d7a9..ad902332 100644
+--- a/sos/report/plugins/pcp.py
++++ b/sos/report/plugins/pcp.py
+@@ -41,7 +41,7 @@ class Pcp(Plugin, RedHatPlugin, DebianPlugin):
+ total_size = 0
+ for dirpath, dirnames, filenames in os.walk(path):
+ for f in filenames:
+- fp = os.path.join(dirpath, f)
++ fp = self.path_join(dirpath, f)
+ total_size += os.path.getsize(fp)
+ return total_size
+
+@@ -86,7 +86,7 @@ class Pcp(Plugin, RedHatPlugin, DebianPlugin):
+ # unconditionally. Obviously if someone messes up their /etc/pcp.conf
+ # in a ridiculous way (i.e. setting PCP_SYSCONF_DIR to '/') this will
+ # break badly.
+- var_conf_dir = os.path.join(self.pcp_var_dir, 'config')
++ var_conf_dir = self.path_join(self.pcp_var_dir, 'config')
+ self.add_copy_spec([
+ self.pcp_sysconf_dir,
+ self.pcp_conffile,
+@@ -98,10 +98,10 @@ class Pcp(Plugin, RedHatPlugin, DebianPlugin):
+ # rpms. It does not make up for a lot of size but it contains many
+ # files
+ self.add_forbidden_path([
+- os.path.join(var_conf_dir, 'pmchart'),
+- os.path.join(var_conf_dir, 'pmlogconf'),
+- os.path.join(var_conf_dir, 'pmieconf'),
+- os.path.join(var_conf_dir, 'pmlogrewrite')
++ self.path_join(var_conf_dir, 'pmchart'),
++ self.path_join(var_conf_dir, 'pmlogconf'),
++ self.path_join(var_conf_dir, 'pmieconf'),
++ self.path_join(var_conf_dir, 'pmlogrewrite')
+ ])
+
+ # Take PCP_LOG_DIR/pmlogger/`hostname` + PCP_LOG_DIR/pmmgr/`hostname`
+@@ -121,13 +121,13 @@ class Pcp(Plugin, RedHatPlugin, DebianPlugin):
+ # we would collect everything
+ if self.pcp_hostname != '':
+ # collect pmmgr logs up to 'pmmgrlogs' size limit
+- path = os.path.join(self.pcp_log_dir, 'pmmgr',
+- self.pcp_hostname, '*')
++ path = self.path_join(self.pcp_log_dir, 'pmmgr',
++ self.pcp_hostname, '*')
+ self.add_copy_spec(path, sizelimit=self.sizelimit, tailit=False)
+ # collect newest pmlogger logs up to 'pmloggerfiles' count
+ files_collected = 0
+- path = os.path.join(self.pcp_log_dir, 'pmlogger',
+- self.pcp_hostname, '*')
++ path = self.path_join(self.pcp_log_dir, 'pmlogger',
++ self.pcp_hostname, '*')
+ pmlogger_ls = self.exec_cmd("ls -t1 %s" % path)
+ if pmlogger_ls['status'] == 0:
+ for line in pmlogger_ls['output'].splitlines():
+@@ -138,15 +138,15 @@ class Pcp(Plugin, RedHatPlugin, DebianPlugin):
+
+ self.add_copy_spec([
+ # Collect PCP_LOG_DIR/pmcd and PCP_LOG_DIR/NOTICES
+- os.path.join(self.pcp_log_dir, 'pmcd'),
+- os.path.join(self.pcp_log_dir, 'NOTICES*'),
++ self.path_join(self.pcp_log_dir, 'pmcd'),
++ self.path_join(self.pcp_log_dir, 'NOTICES*'),
+ # Collect PCP_VAR_DIR/pmns
+- os.path.join(self.pcp_var_dir, 'pmns'),
++ self.path_join(self.pcp_var_dir, 'pmns'),
+ # Also collect any other log and config files
+ # (as suggested by fche)
+- os.path.join(self.pcp_log_dir, '*/*.log*'),
+- os.path.join(self.pcp_log_dir, '*/*/*.log*'),
+- os.path.join(self.pcp_log_dir, '*/*/config*')
++ self.path_join(self.pcp_log_dir, '*/*.log*'),
++ self.path_join(self.pcp_log_dir, '*/*/*.log*'),
++ self.path_join(self.pcp_log_dir, '*/*/config*')
+ ])
+
+ # Collect a summary for the current day
+diff --git a/sos/report/plugins/postfix.py b/sos/report/plugins/postfix.py
+index 8f584430..3ca0c4ad 100644
+--- a/sos/report/plugins/postfix.py
++++ b/sos/report/plugins/postfix.py
+@@ -41,7 +41,7 @@ class Postfix(Plugin):
+ ]
+ fp = []
+ try:
+- with open('/etc/postfix/main.cf', 'r') as cffile:
++ with open(self.path_join('/etc/postfix/main.cf'), 'r') as cffile:
+ for line in cffile.readlines():
+ # ignore comments and take the first word after '='
+ if line.startswith('#'):
+diff --git a/sos/report/plugins/postgresql.py b/sos/report/plugins/postgresql.py
+index bec0b019..00824db7 100644
+--- a/sos/report/plugins/postgresql.py
++++ b/sos/report/plugins/postgresql.py
+@@ -124,7 +124,7 @@ class RedHatPostgreSQL(PostgreSQL, SCLPlugin):
+
+ # copy PG_VERSION and postmaster.opts
+ for f in ["PG_VERSION", "postmaster.opts"]:
+- self.add_copy_spec(os.path.join(_dir, "data", f))
++ self.add_copy_spec(self.path_join(_dir, "data", f))
+
+
+ class DebianPostgreSQL(PostgreSQL, DebianPlugin, UbuntuPlugin):
+diff --git a/sos/report/plugins/powerpc.py b/sos/report/plugins/powerpc.py
+index 4fb4f87c..50f88650 100644
+--- a/sos/report/plugins/powerpc.py
++++ b/sos/report/plugins/powerpc.py
+@@ -22,7 +22,7 @@ class PowerPC(Plugin, IndependentPlugin):
+
+ def setup(self):
+ try:
+- with open('/proc/cpuinfo', 'r') as fp:
++ with open(self.path_join('/proc/cpuinfo'), 'r') as fp:
+ contents = fp.read()
+ ispSeries = "pSeries" in contents
+ isPowerNV = "PowerNV" in contents
+diff --git a/sos/report/plugins/processor.py b/sos/report/plugins/processor.py
+index 2df2dc9a..c3d8930c 100644
+--- a/sos/report/plugins/processor.py
++++ b/sos/report/plugins/processor.py
+@@ -7,7 +7,6 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ from sos.report.plugins import Plugin, IndependentPlugin
+-import os
+
+
+ class Processor(Plugin, IndependentPlugin):
+@@ -41,7 +40,7 @@ class Processor(Plugin, IndependentPlugin):
+ # cumulative directory size exceeds 25MB or even 100MB.
+ cdirs = self.listdir('/sys/devices/system/cpu')
+ self.add_copy_spec([
+- os.path.join('/sys/devices/system/cpu', cdir) for cdir in cdirs
++ self.path_join('/sys/devices/system/cpu', cdir) for cdir in cdirs
+ ])
+
+ self.add_cmd_output([
+diff --git a/sos/report/plugins/python.py b/sos/report/plugins/python.py
+index e2ab39ab..a8ec0cd8 100644
+--- a/sos/report/plugins/python.py
++++ b/sos/report/plugins/python.py
+@@ -68,9 +68,9 @@ class RedHatPython(Python, RedHatPlugin):
+ ]
+
+ for py_path in py_paths:
+- for root, _, files in os.walk(py_path):
++ for root, _, files in os.walk(self.path_join(py_path)):
+ for file_ in files:
+- filepath = os.path.join(root, file_)
++ filepath = self.path_join(root, file_)
+ if filepath.endswith('.py'):
+ try:
+ with open(filepath, 'rb') as f:
+diff --git a/sos/report/plugins/sar.py b/sos/report/plugins/sar.py
+index 669f5d7b..b60005b1 100644
+--- a/sos/report/plugins/sar.py
++++ b/sos/report/plugins/sar.py
+@@ -7,7 +7,6 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ from sos.report.plugins import Plugin, RedHatPlugin, DebianPlugin, UbuntuPlugin
+-import os
+ import re
+
+
+@@ -24,7 +23,7 @@ class Sar(Plugin,):
+ "", False)]
+
+ def setup(self):
+- self.add_copy_spec(os.path.join(self.sa_path, '*'),
++ self.add_copy_spec(self.path_join(self.sa_path, '*'),
+ sizelimit=0 if self.get_option("all_sar") else None,
+ tailit=False)
+
+@@ -44,7 +43,7 @@ class Sar(Plugin,):
+ # as option for sadc
+ for fname in dir_list:
+ if sa_regex.match(fname):
+- sa_data_path = os.path.join(self.sa_path, fname)
++ sa_data_path = self.path_join(self.sa_path, fname)
+ sar_filename = 'sar' + fname[2:]
+ if sar_filename not in dir_list:
+ sar_cmd = 'sh -c "sar -A -f %s"' % sa_data_path
+diff --git a/sos/report/plugins/sos_extras.py b/sos/report/plugins/sos_extras.py
+index ffde4138..55bc4dc0 100644
+--- a/sos/report/plugins/sos_extras.py
++++ b/sos/report/plugins/sos_extras.py
+@@ -58,7 +58,7 @@ class SosExtras(Plugin, IndependentPlugin):
+
+ for path, dirlist, filelist in os.walk(self.extras_dir):
+ for f in filelist:
+- _file = os.path.join(path, f)
++ _file = self.path_join(path, f)
+ self._log_warn("Collecting data from extras file %s" % _file)
+ try:
+ for line in open(_file).read().splitlines():
+diff --git a/sos/report/plugins/ssh.py b/sos/report/plugins/ssh.py
+index 971cda8b..9ac9dec0 100644
+--- a/sos/report/plugins/ssh.py
++++ b/sos/report/plugins/ssh.py
+@@ -42,7 +41,7 @@ class Ssh(Plugin, IndependentPlugin):
+ try:
+ for sshcfg in sshcfgs:
+ tag = sshcfg.split('/')[-1]
+- with open(sshcfg, 'r') as cfgfile:
++ with open(self.path_join(sshcfg), 'r') as cfgfile:
+ for line in cfgfile:
+ # skip empty lines and comments
+ if len(line.split()) == 0 or line.startswith('#'):
+diff --git a/sos/report/plugins/unpackaged.py b/sos/report/plugins/unpackaged.py
+index 9205e53f..772b1d1f 100644
+--- a/sos/report/plugins/unpackaged.py
++++ b/sos/report/plugins/unpackaged.py
+@@ -40,7 +40,7 @@ class Unpackaged(Plugin, RedHatPlugin):
+ for e in exclude:
+ dirs[:] = [d for d in dirs if d not in e]
+ for name in files:
+- path = os.path.join(root, name)
++ path = self.path_join(root, name)
+ try:
+ if stat.S_ISLNK(os.lstat(path).st_mode):
+ path = Path(path).resolve()
+@@ -49,7 +49,7 @@ class Unpackaged(Plugin, RedHatPlugin):
+ file_list.append(os.path.realpath(path))
+ for name in dirs:
+ file_list.append(os.path.realpath(
+- os.path.join(root, name)))
++ self.path_join(root, name)))
+
+ return file_list
+
+diff --git a/sos/report/plugins/watchdog.py b/sos/report/plugins/watchdog.py
+index 1bf3f4cb..bf2dc9cb 100644
+--- a/sos/report/plugins/watchdog.py
++++ b/sos/report/plugins/watchdog.py
+@@ -11,7 +11,6 @@
+ from sos.report.plugins import Plugin, RedHatPlugin
+
+ from glob import glob
+-import os
+
+
+ class Watchdog(Plugin, RedHatPlugin):
+@@ -56,8 +55,8 @@ class Watchdog(Plugin, RedHatPlugin):
+ Collect configuration files, custom executables for test-binary
+ and repair-binary, and stdout/stderr logs.
+ """
+- conf_file = self.get_option('conf_file')
+- log_dir = '/var/log/watchdog'
++ conf_file = self.path_join(self.get_option('conf_file'))
++ log_dir = self.path_join('/var/log/watchdog')
+
+ # Get service configuration and sysconfig files
+ self.add_copy_spec([
+@@ -80,15 +79,15 @@ class Watchdog(Plugin, RedHatPlugin):
+ self._log_warn("Could not read %s: %s" % (conf_file, ex))
+
+ if self.get_option('all_logs'):
+- log_files = glob(os.path.join(log_dir, '*'))
++ log_files = glob(self.path_join(log_dir, '*'))
+ else:
+- log_files = (glob(os.path.join(log_dir, '*.stdout')) +
+- glob(os.path.join(log_dir, '*.stderr')))
++ log_files = (glob(self.path_join(log_dir, '*.stdout')) +
++ glob(self.path_join(log_dir, '*.stderr')))
+
+ self.add_copy_spec(log_files)
+
+ # Get output of "wdctl <device>" for each /dev/watchdog*
+- for dev in glob('/dev/watchdog*'):
++ for dev in glob(self.path_join('/dev/watchdog*')):
+ self.add_cmd_output("wdctl %s" % dev)
+
+ # vim: set et ts=4 sw=4 :
+diff --git a/sos/report/plugins/yum.py b/sos/report/plugins/yum.py
+index 148464cb..e5256642 100644
+--- a/sos/report/plugins/yum.py
++++ b/sos/report/plugins/yum.py
+@@ -61,7 +61,7 @@ class Yum(Plugin, RedHatPlugin):
+ if not p.endswith(".py"):
+ continue
+ plugins = plugins + " " if len(plugins) else ""
+- plugins = plugins + os.path.join(YUM_PLUGIN_PATH, p)
++ plugins = plugins + self.path_join(YUM_PLUGIN_PATH, p)
+ if len(plugins):
+ self.add_cmd_output("rpm -qf %s" % plugins,
+ suggest_filename="plugin-packages")
+--
+2.31.1
+
+From f4af5efdc79aefe1aa685c36d095925bae14dc4a Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 28 Sep 2021 13:00:17 -0400
+Subject: [PATCH 1/4] [collect] Add --transport option and allow clusters to
+ set transport type
+
+Adds a new `--transport` option for users to be able to specify the type
+of transport to use when connecting to nodes. The default value of
+`auto` will defer to the cluster profile to set the transport type,
+which will continue to default to use OpenSSH's ControlPersist feature.
+
+Clusters may override the new `set_transport_type()` method to change
+the default transport used.
+
+If `--transport` is anything besides `auto`, then the cluster profile
+will not be deferred to when choosing a transport for each remote node.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ man/en/sos-collect.1 | 15 +++++++++++++++
+ sos/collector/__init__.py | 6 ++++++
+ sos/collector/clusters/__init__.py | 10 ++++++++++
+ sos/collector/exceptions.py | 13 ++++++++++++-
+ sos/collector/sosnode.py | 16 +++++++++++++++-
+ 5 files changed, 58 insertions(+), 2 deletions(-)
+
+diff --git a/man/en/sos-collect.1 b/man/en/sos-collect.1
+index e930023e..8ad4fe5e 100644
+--- a/man/en/sos-collect.1
++++ b/man/en/sos-collect.1
+@@ -43,6 +43,7 @@ sos collect \- Collect sosreports from multiple (cluster) nodes
+ [\-\-sos-cmd SOS_CMD]
+ [\-t|\-\-threads THREADS]
+ [\-\-timeout TIMEOUT]
++ [\-\-transport TRANSPORT]
+ [\-\-tmp\-dir TMP_DIR]
+ [\-v|\-\-verbose]
+ [\-\-verify]
+@@ -350,6 +351,20 @@ Note that sosreports are collected in parallel, so you can approximate the total
+ runtime of sos collect via timeout*(number of nodes/jobs).
+
+ Default is 180 seconds.
++.TP
++\fB\-\-transport\fR TRANSPORT
++Specify the type of remote transport to use to manage connections to remote nodes.
++
++\fBsos collect\fR uses locally installed binaries to connect to and interact with remote
++nodes, instead of directly establishing those connections. By default, OpenSSH's ControlPersist
++feature is preferred, however certain cluster types may have preferences of their own for how
++remote sessions should be established.
++
++The types of transports supported are currently as follows:
++
++ \fBauto\fR Allow the cluster type to determine the transport used
++ \fBcontrol_persist\fR Use OpenSSH's ControlPersist feature. This is the default behavior
++
+ .TP
+ \fB\-\-tmp\-dir\fR TMP_DIR
+ Specify a temporary directory to save sos archives to. By default one will be created in
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index da912655..fecfe6aa 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -98,6 +98,7 @@ class SoSCollector(SoSComponent):
+ 'ssh_port': 22,
+ 'ssh_user': 'root',
+ 'timeout': 600,
++ 'transport': 'auto',
+ 'verify': False,
+ 'usernames': [],
+ 'upload': False,
+@@ -378,6 +379,8 @@ class SoSCollector(SoSComponent):
+ help='Specify an SSH user. Default root')
+ collect_grp.add_argument('--timeout', type=int, required=False,
+ help='Timeout for sosreport on each node.')
++ collect_grp.add_argument('--transport', default='auto', type=str,
++ help='Remote connection transport to use')
+ collect_grp.add_argument("--upload", action="store_true",
+ default=False,
+ help="Upload archive to a policy-default "
+@@ -813,6 +813,8 @@ class SoSCollector(SoSComponent):
+ self.collect_md.add_field('cluster_type', self.cluster_type)
+ if self.cluster:
+ self.master.cluster = self.cluster
++ if self.opts.transport == 'auto':
++ self.opts.transport = self.cluster.set_transport_type()
+ self.cluster.setup()
+ if self.cluster.cluster_ssh_key:
+ if not self.opts.ssh_key:
+@@ -1041,6 +1046,7 @@ class SoSCollector(SoSComponent):
+ else:
+ client.disconnect()
+ except Exception:
++ # all exception logging is handled within SoSNode
+ pass
+
+ def intro(self):
+diff --git a/sos/collector/clusters/__init__.py b/sos/collector/clusters/__init__.py
+index 64ac2a44..cf1e7a0b 100644
+--- a/sos/collector/clusters/__init__.py
++++ b/sos/collector/clusters/__init__.py
+@@ -149,6 +149,16 @@ class Cluster():
+ """
+ pass
+
++ def set_transport_type(self):
++ """The default connection type used by sos collect is to leverage the
++ local system's SSH installation using ControlPersist, however certain
++ cluster types may want to use something else.
++
++ Override this in a specific cluster profile to set the ``transport``
++ option according to what type of transport should be used.
++ """
++ return 'control_persist'
++
+ def set_master_options(self, node):
+ """If there is a need to set specific options in the sos command being
+ run on the cluster's master nodes, override this method in the cluster
+diff --git a/sos/collector/exceptions.py b/sos/collector/exceptions.py
+index 1e44768b..2bb07e7b 100644
+--- a/sos/collector/exceptions.py
++++ b/sos/collector/exceptions.py
+@@ -94,6 +94,16 @@ class UnsupportedHostException(Exception):
+ super(UnsupportedHostException, self).__init__(message)
+
+
++class InvalidTransportException(Exception):
++ """Raised when a transport is requested but it does not exist or is
++ not supported locally"""
++
++ def __init__(self, transport=None):
++ message = ("Connection failed: unknown or unsupported transport %s"
++ % transport if transport else '')
++ super(InvalidTransportException, self).__init__(message)
++
++
+ __all__ = [
+ 'AuthPermissionDeniedException',
+ 'CommandTimeoutException',
+@@ -104,5 +114,6 @@ __all__ = [
+ 'InvalidPasswordException',
+ 'PasswordRequestException',
+ 'TimeoutPasswordAuthException',
+- 'UnsupportedHostException'
++ 'UnsupportedHostException',
++ 'InvalidTransportException'
+ ]
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index f79bd5ff..5c5c7201 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -22,7 +22,13 @@ from sos.collector.transports.control_persist import SSHControlPersist
+ from sos.collector.transports.local import LocalTransport
+ from sos.collector.exceptions import (CommandTimeoutException,
+ ConnectionException,
+- UnsupportedHostException)
++ UnsupportedHostException,
++ InvalidTransportException)
++
++TRANSPORTS = {
++ 'local': LocalTransport,
++ 'control_persist': SSHControlPersist,
++}
+
+
+ class SosNode():
+@@ -107,6 +113,14 @@ class SosNode():
+ if self.address in ['localhost', '127.0.0.1']:
+ self.local = True
+ return LocalTransport(self.address, commons)
++ elif self.opts.transport in TRANSPORTS.keys():
++ return TRANSPORTS[self.opts.transport](self.address, commons)
++ elif self.opts.transport != 'auto':
++ self.log_error(
++ "Connection failed: unknown or unsupported transport %s"
++ % self.opts.transport
++ )
++ raise InvalidTransportException(self.opts.transport)
+ return SSHControlPersist(self.address, commons)
+
+ def _fmt_msg(self, msg):
+--
+2.31.1
+
+
+From dbc49345384404600f45d68b8d3c6541b2a26480 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 30 Sep 2021 10:38:18 -0400
+Subject: [PATCH 2/4] [transports] Add 'oc' as a transport option for remote
+ nodes
+
+This commit adds a new transport for `sos collect` by leveraging a
+locally available `oc` binary that has been properly configured for
+access to an OCP cluster.
+
+This transport will allow users to use `sos collect` to collect reports
+from an OCP cluster without directly connecting to any of the nodes
+involved. We do this by using the `oc` binary to first launch a pod on
+target node(s) and then exec our discovery commands and eventual `sos
+report` command to that pod. This in turn is dependent on a function API
+for the `oc` binary to communicate with. In the event that `oc` is not
+__locally__ available or is not properly configured, we will fallback to
+the current default of using SSH ControlPersist to directly connect to
+the nodes. Otherwise, the OCP cluster will attempt to automatically use
+this new transport.
+---
+ man/en/sos-collect.1 | 1 +
+ sos/collector/clusters/ocp.py | 14 ++
+ sos/collector/sosnode.py | 8 +-
+ sos/collector/transports/__init__.py | 20 ++-
+ sos/collector/transports/oc.py | 220 +++++++++++++++++++++++++++
+ 5 files changed, 257 insertions(+), 6 deletions(-)
+ create mode 100644 sos/collector/transports/oc.py
+
+diff --git a/man/en/sos-collect.1 b/man/en/sos-collect.1
+index 8ad4fe5e..a1f6c10e 100644
+--- a/man/en/sos-collect.1
++++ b/man/en/sos-collect.1
+@@ -364,6 +364,7 @@ The types of transports supported are currently as follows:
+
+ \fBauto\fR Allow the cluster type to determine the transport used
+ \fBcontrol_persist\fR Use OpenSSH's ControlPersist feature. This is the default behavior
++ \fBoc\fR Use a \fBlocally\fR configured \fBoc\fR binary to deploy collection pods on OCP nodes
+
+ .TP
+ \fB\-\-tmp\-dir\fR TMP_DIR
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index ad97587f..a9357dbf 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -12,6 +12,7 @@ import os
+
+ from pipes import quote
+ from sos.collector.clusters import Cluster
++from sos.utilities import is_executable
+
+
+ class ocp(Cluster):
+@@ -83,6 +84,19 @@ class ocp(Cluster):
+ nodes[_node[0]][column] = _node[idx[column]]
+ return nodes
+
++ def set_transport_type(self):
++ if is_executable('oc'):
++ return 'oc'
++ self.log_info("Local installation of 'oc' not found or is not "
++ "correctly configured. Will use ControlPersist")
++ self.ui_log.warn(
++ "Preferred transport 'oc' not available, will fallback to SSH."
++ )
++ if not self.opts.batch:
++ input("Press ENTER to continue connecting with SSH, or Ctrl+C to"
++ "abort.")
++ return 'control_persist'
++
+ def get_nodes(self):
+ nodes = []
+ self.node_dict = {}
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 5c5c7201..8a9dbd7a 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -20,6 +20,7 @@ from sos.policies import load
+ from sos.policies.init_systems import InitSystem
+ from sos.collector.transports.control_persist import SSHControlPersist
+ from sos.collector.transports.local import LocalTransport
++from sos.collector.transports.oc import OCTransport
+ from sos.collector.exceptions import (CommandTimeoutException,
+ ConnectionException,
+ UnsupportedHostException,
+@@ -28,6 +29,7 @@ from sos.collector.exceptions import (CommandTimeoutException,
+ TRANSPORTS = {
+ 'local': LocalTransport,
+ 'control_persist': SSHControlPersist,
++ 'oc': OCTransport
+ }
+
+
+@@ -421,13 +423,11 @@ class SosNode():
+ if 'atomic' in cmd:
+ get_pty = True
+
+- if get_pty:
+- cmd = "/bin/bash -c %s" % quote(cmd)
+-
+ if env:
+ _cmd_env = self.env_vars
+ env = _cmd_env.update(env)
+- return self._transport.run_command(cmd, timeout, need_root, env)
++ return self._transport.run_command(cmd, timeout, need_root, env,
++ get_pty)
+
+ def sosreport(self):
+ """Run an sos report on the node, then collect it"""
+diff --git a/sos/collector/transports/__init__.py b/sos/collector/transports/__init__.py
+index 5be7dc6d..7bffee62 100644
+--- a/sos/collector/transports/__init__.py
++++ b/sos/collector/transports/__init__.py
+@@ -144,7 +144,8 @@ class RemoteTransport():
+ raise NotImplementedError("Transport %s does not define disconnect"
+ % self.name)
+
+- def run_command(self, cmd, timeout=180, need_root=False, env=None):
++ def run_command(self, cmd, timeout=180, need_root=False, env=None,
++ get_pty=False):
+ """Run a command on the node, returning its output and exit code.
+ This should return the exit code of the command being executed, not the
+ exit code of whatever mechanism the transport uses to execute that
+@@ -165,10 +166,15 @@ class RemoteTransport():
+ :param env: Specify env vars to be passed to the ``cmd``
+ :type env: ``dict``
+
++ :param get_pty: Does ``cmd`` require execution with a pty?
++ :type get_pty: ``bool``
++
+ :returns: Output of ``cmd`` and the exit code
+ :rtype: ``dict`` with keys ``output`` and ``status``
+ """
+ self.log_debug('Running command %s' % cmd)
++ if get_pty:
++ cmd = "/bin/bash -c %s" % quote(cmd)
+ # currently we only use/support the use of pexpect for handling the
+ # execution of these commands, as opposed to directly invoking
+ # subprocess.Popen() in conjunction with tools like sshpass.
+@@ -212,6 +218,13 @@ class RemoteTransport():
+ :type env: ``dict``
+ """
+ cmd = self._format_cmd_for_exec(cmd)
++
++ # if for any reason env is empty, set it to None as otherwise
++ # pexpect interprets this to mean "run this command with no env vars of
++ # any kind"
++ if not env:
++ env = None
++
+ result = pexpect.spawn(cmd, encoding='utf-8', env=env)
+
+ _expects = [pexpect.EOF, pexpect.TIMEOUT]
+@@ -268,6 +281,9 @@ class RemoteTransport():
+ _out = self.run_command('hostname')
+ if _out['status'] == 0:
+ self._hostname = _out['output'].strip()
++
++ if not self._hostname:
++ self._hostname = self.address
+ self.log_info("Hostname set to %s" % self._hostname)
+ return self._hostname
+
+@@ -302,7 +318,7 @@ class RemoteTransport():
+ return self._read_file(fname)
+
+ def _read_file(self, fname):
+- res = self.run_command("cat %s" % fname, timeout=5)
++ res = self.run_command("cat %s" % fname, timeout=10)
+ if res['status'] == 0:
+ return res['output']
+ else:
+diff --git a/sos/collector/transports/oc.py b/sos/collector/transports/oc.py
+new file mode 100644
+index 00000000..649037b9
+--- /dev/null
++++ b/sos/collector/transports/oc.py
+@@ -0,0 +1,220 @@
++# Copyright Red Hat 2021, Jake Hunsaker <jhunsake@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++import json
++import tempfile
++import os
++
++from sos.collector.transports import RemoteTransport
++from sos.utilities import (is_executable, sos_get_command_output,
++ SoSTimeoutError)
++
++
++class OCTransport(RemoteTransport):
++ """This transport leverages the execution of commands via a locally
++ available and configured ``oc`` binary for OCPv4 environments.
++
++ OCPv4 clusters generally discourage the use of SSH, so this transport may
++ be used to remove our use of SSH in favor of the environment provided
++ method of connecting to nodes and executing commands via debug pods.
++
++ Note that this approach will generate multiple debug pods over the course
++ of our execution
++ """
++
++ name = 'oc'
++ project = 'sos-collect-tmp'
++
++ def run_oc(self, cmd, **kwargs):
++ """Format and run a command with `oc` in the project defined for our
++ execution
++ """
++ return sos_get_command_output(
++ "oc -n sos-collect-tmp %s" % cmd,
++ **kwargs
++ )
++
++ @property
++ def connected(self):
++ up = self.run_oc(
++ "wait --timeout=0s --for=condition=ready pod/%s" % self.pod_name
++ )
++ return up['status'] == 0
++
++ def get_node_pod_config(self):
++ """Based on our template for the debug container, add the node-specific
++ items so that we can deploy one of these on each node we're collecting
++ from
++ """
++ return {
++ "kind": "Pod",
++ "apiVersion": "v1",
++ "metadata": {
++ "name": "%s-sos-collector" % self.address.split('.')[0],
++ "namespace": "sos-collect-tmp"
++ },
++ "priorityClassName": "system-cluster-critical",
++ "spec": {
++ "volumes": [
++ {
++ "name": "host",
++ "hostPath": {
++ "path": "/",
++ "type": "Directory"
++ }
++ },
++ {
++ "name": "run",
++ "hostPath": {
++ "path": "/run",
++ "type": "Directory"
++ }
++ },
++ {
++ "name": "varlog",
++ "hostPath": {
++ "path": "/var/log",
++ "type": "Directory"
++ }
++ },
++ {
++ "name": "machine-id",
++ "hostPath": {
++ "path": "/etc/machine-id",
++ "type": "File"
++ }
++ }
++ ],
++ "containers": [
++ {
++ "name": "sos-collector-tmp",
++ "image": "registry.redhat.io/rhel8/support-tools",
++ "command": [
++ "/bin/bash"
++ ],
++ "env": [
++ {
++ "name": "HOST",
++ "value": "/host"
++ }
++ ],
++ "resources": {},
++ "volumeMounts": [
++ {
++ "name": "host",
++ "mountPath": "/host"
++ },
++ {
++ "name": "run",
++ "mountPath": "/run"
++ },
++ {
++ "name": "varlog",
++ "mountPath": "/var/log"
++ },
++ {
++ "name": "machine-id",
++ "mountPath": "/etc/machine-id"
++ }
++ ],
++ "securityContext": {
++ "privileged": True,
++ "runAsUser": 0
++ },
++ "stdin": True,
++ "stdinOnce": True,
++ "tty": True
++ }
++ ],
++ "restartPolicy": "Never",
++ "nodeName": self.address,
++ "hostNetwork": True,
++ "hostPID": True,
++ "hostIPC": True
++ }
++ }
++
++ def _connect(self, password):
++ # the oc binary must be _locally_ available for this to work
++ if not is_executable('oc'):
++ return False
++
++ # deploy the debug container we'll exec into
++ podconf = self.get_node_pod_config()
++ self.pod_name = podconf['metadata']['name']
++ fd, self.pod_tmp_conf = tempfile.mkstemp(dir=self.tmpdir)
++ with open(fd, 'w') as cfile:
++ json.dump(podconf, cfile)
++ self.log_debug("Starting sos collector container '%s'" % self.pod_name)
++ # this specifically does not need to run with a project definition
++ out = sos_get_command_output(
++ "oc create -f %s" % self.pod_tmp_conf
++ )
++ if (out['status'] != 0 or "pod/%s created" % self.pod_name not in
++ out['output']):
++ self.log_error("Unable to deploy sos collect pod")
++ self.log_debug("Debug pod deployment failed: %s" % out['output'])
++ return False
++ self.log_debug("Pod '%s' successfully deployed, waiting for pod to "
++ "enter ready state" % self.pod_name)
++
++ # wait for the pod to report as running
++ try:
++ up = self.run_oc("wait --for=condition=Ready pod/%s --timeout=30s"
++ % self.pod_name,
++ # timeout is for local safety, not oc
++ timeout=40)
++ if not up['status'] == 0:
++ self.log_error("Pod not available after 30 seconds")
++ return False
++ except SoSTimeoutError:
++ self.log_error("Timeout while polling for pod readiness")
++ return False
++ except Exception as err:
++ self.log_error("Error while waiting for pod to be ready: %s"
++ % err)
++ return False
++
++ return True
++
++ def _format_cmd_for_exec(self, cmd):
++ if cmd.startswith('oc'):
++ return ("oc -n %s exec --request-timeout=0 %s -- chroot /host %s"
++ % (self.project, self.pod_name, cmd))
++ return super(OCTransport, self)._format_cmd_for_exec(cmd)
++
++ def run_command(self, cmd, timeout=180, need_root=False, env=None,
++ get_pty=False):
++ # debug pod setup is slow, extend all timeouts to account for this
++ if timeout:
++ timeout += 10
++
++ # since we always execute within a bash shell, force disable get_pty
++ # to avoid double-quoting
++ return super(OCTransport, self).run_command(cmd, timeout, need_root,
++ env, False)
++
++ def _disconnect(self):
++ os.unlink(self.pod_tmp_conf)
++ removed = self.run_oc("delete pod %s" % self.pod_name)
++ if "deleted" not in removed['output']:
++ self.log_debug("Calling delete on pod '%s' failed: %s"
++ % (self.pod_name, removed))
++ return False
++ return True
++
++ @property
++ def remote_exec(self):
++ return ("oc -n %s exec --request-timeout=0 %s -- /bin/bash -c"
++ % (self.project, self.pod_name))
++
++ def _retrieve_file(self, fname, dest):
++ cmd = self.run_oc("cp %s:%s %s" % (self.pod_name, fname, dest))
++ return cmd['status'] == 0
+--
+2.31.1
+
+
+From 460494c4296db1a7529b44fe8f6597544c917c02 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 11 Oct 2021 11:50:44 -0400
+Subject: [PATCH 3/4] [ocp] Create temporary project and restrict default node
+ list to masters
+
+Adds explicit setup of a new project to use in the `ocp` cluster and
+adds better handling of cluster setup generally, which the `ocp` cluster
+is the first to make use of.
+
+Included in this change is a correction to
+`Cluster.exec_primary_cmd()`'s use of `get_pty` to now be determined on
+if the primary node is the local node or not.
+
+Additionally, based on feedback from the OCP engineering team, by
+default restrict node lists to masters.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/__init__.py | 5 ++++
+ sos/collector/clusters/__init__.py | 13 +++++++-
+ sos/collector/clusters/ocp.py | 48 ++++++++++++++++++++++++++++--
+ sos/collector/transports/oc.py | 4 +--
+ 4 files changed, 64 insertions(+), 6 deletions(-)
+
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index fecfe6aa..a76f8a79 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -850,6 +850,7 @@ class SoSCollector(SoSComponent):
+ "CTRL-C to quit\n")
+ self.ui_log.info("")
+ except KeyboardInterrupt:
++ self.cluster.cleanup()
+ self.exit("Exiting on user cancel", 130)
+
+ def configure_sos_cmd(self):
+@@ -1098,6 +1099,7 @@ this utility or remote systems that it connects to.
+ self.archive.makedirs('sos_logs', 0o755)
+
+ self.collect()
++ self.cluster.cleanup()
+ self.cleanup()
+
+ def collect(self):
+@@ -1156,9 +1158,11 @@ this utility or remote systems that it connects to.
+ pool.shutdown(wait=True)
+ except KeyboardInterrupt:
+ self.log_error('Exiting on user cancel\n')
++ self.cluster.cleanup()
+ os._exit(130)
+ except Exception as err:
+ self.log_error('Could not connect to nodes: %s' % err)
++ self.cluster.cleanup()
+ os._exit(1)
+
+ if hasattr(self.cluster, 'run_extra_cmd'):
+@@ -1199,6 +1199,7 @@ this utility or remote systems that it c
+ arc_name = self.create_cluster_archive()
+ else:
+ msg = 'No sosreports were collected, nothing to archive...'
++ self.cluster.cleanup()
+ self.exit(msg, 1)
+
+ if self.opts.upload and self.policy.get_upload_url():
+diff --git a/sos/collector/clusters/__init__.py b/sos/collector/clusters/__init__.py
+index cf1e7a0b..2a4665a1 100644
+--- a/sos/collector/clusters/__init__.py
++++ b/sos/collector/clusters/__init__.py
+@@ -192,7 +192,8 @@ class Cluster():
+ :returns: The output and status of `cmd`
+ :rtype: ``dict``
+ """
+- res = self.master.run_command(cmd, get_pty=True, need_root=need_root)
++ pty = self.master.local is False
++ res = self.master.run_command(cmd, get_pty=pty, need_root=need_root)
+ if res['output']:
+ res['output'] = res['output'].replace('Password:', '')
+ return res
+@@ -223,6 +224,16 @@ class Cluster():
+ return True
+ return False
+
++ def cleanup(self):
++ """
++ This may be overridden by clusters
++
++ Perform any necessary cleanup steps required by the cluster profile.
++ This helps ensure that sos does make lasting changes to the environment
++ in which we are running
++ """
++ pass
++
+ def get_nodes(self):
+ """
+ This MUST be overridden by a cluster profile subclassing this class
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index a9357dbf..92da4e6e 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -23,10 +23,12 @@ class ocp(Cluster):
+
+ api_collect_enabled = False
+ token = None
++ project = 'sos-collect-tmp'
++ oc_cluster_admin = None
+
+ option_list = [
+ ('label', '', 'Colon delimited list of labels to select nodes with'),
+- ('role', '', 'Colon delimited list of roles to select nodes with'),
++ ('role', 'master', 'Colon delimited list of roles to filter on'),
+ ('kubeconfig', '', 'Path to the kubeconfig file'),
+ ('token', '', 'Service account token to use for oc authorization')
+ ]
+@@ -58,6 +58,42 @@ class ocp(Cluster):
+ _who = self.fmt_oc_cmd('whoami')
+ return self.exec_master_cmd(_who)['status'] == 0
+
++ def setup(self):
++ """Create the project that we will be executing in for any nodes'
++ collection via a container image
++ """
++ if not self.set_transport_type() == 'oc':
++ return
++
++ out = self.exec_master_cmd(self.fmt_oc_cmd("auth can-i '*' '*'"))
++ self.oc_cluster_admin = out['status'] == 0
++ if not self.oc_cluster_admin:
++ self.log_debug("Check for cluster-admin privileges returned false,"
++ " cannot create project in OCP cluster")
++ raise Exception("Insufficient permissions to create temporary "
++ "collection project.\nAborting...")
++
++ self.log_info("Creating new temporary project '%s'" % self.project)
++ ret = self.exec_master_cmd("oc new-project %s" % self.project)
++ if ret['status'] == 0:
++ return True
++
++ self.log_debug("Failed to create project: %s" % ret['output'])
++ raise Exception("Failed to create temporary project for collection. "
++ "\nAborting...")
++
++ def cleanup(self):
++ """Remove the project we created to execute within
++ """
++ if self.project:
++ ret = self.exec_master_cmd("oc delete project %s" % self.project)
++ if not ret['status'] == 0:
++ self.log_error("Error deleting temporary project: %s"
++ % ret['output'])
++ # don't leave the config on a non-existing project
++ self.exec_master_cmd("oc project default")
++ return True
++
+ def _build_dict(self, nodelist):
+ """From the output of get_nodes(), construct an easier-to-reference
+ dict of nodes that will be used in determining labels, master status,
+@@ -85,10 +123,10 @@ class ocp(Cluster):
+ return nodes
+
+ def set_transport_type(self):
+- if is_executable('oc'):
++ if is_executable('oc') or self.opts.transport == 'oc':
+ return 'oc'
+ self.log_info("Local installation of 'oc' not found or is not "
+- "correctly configured. Will use ControlPersist")
++ "correctly configured. Will use ControlPersist.")
+ self.ui_log.warn(
+ "Preferred transport 'oc' not available, will fallback to SSH."
+ )
+@@ -106,6 +144,10 @@ class ocp(Cluster):
+ cmd += " -l %s" % quote(labels)
+ res = self.exec_master_cmd(self.fmt_oc_cmd(cmd))
+ if res['status'] == 0:
++ if self.get_option('role') == 'master':
++ self.log_warn("NOTE: By default, only master nodes are listed."
++ "\nTo collect from all/more nodes, override the "
++ "role option with '-c ocp.role=role1:role2'")
+ roles = [r for r in self.get_option('role').split(':')]
+ self.node_dict = self._build_dict(res['output'].splitlines())
+ for node in self.node_dict:
+diff --git a/sos/collector/transports/oc.py b/sos/collector/transports/oc.py
+index 649037b9..de044ccb 100644
+--- a/sos/collector/transports/oc.py
++++ b/sos/collector/transports/oc.py
+@@ -37,7 +37,7 @@ class OCTransport(RemoteTransport):
+ execution
+ """
+ return sos_get_command_output(
+- "oc -n sos-collect-tmp %s" % cmd,
++ "oc -n %s %s" % (self.project, cmd),
+ **kwargs
+ )
+
+@@ -58,7 +58,7 @@ class OCTransport(RemoteTransport):
+ "apiVersion": "v1",
+ "metadata": {
+ "name": "%s-sos-collector" % self.address.split('.')[0],
+- "namespace": "sos-collect-tmp"
++ "namespace": self.project
+ },
+ "priorityClassName": "system-cluster-critical",
+ "spec": {
+--
+2.31.1
+
+
+From 1bc0e9fe32491e764e622368bfe216f97bf32620 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 4 Oct 2021 15:12:04 -0400
+Subject: [PATCH 4/4] [sosnode] Fix typo and small logic break
+
+Fixes a typo in setting the non-primary node options from the ocp
+profile against the sosnode object. Second, fixes a small break in
+checksum handling for the manifest discovered during `oc` transport
+testing for edge cases.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/clusters/ocp.py | 4 ++--
+ sos/collector/sosnode.py | 4 +++-
+ 2 files changed, 5 insertions(+), 3 deletions(-)
+
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index 92da4e6e..22a7289a 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -183,7 +183,7 @@ class ocp(Cluster):
+ if self.api_collect_enabled:
+ # a primary has already been enabled for API collection, disable
+ # it among others
+- node.plugin_options.append('openshift.no-oc=on')
++ node.plugopts.append('openshift.no-oc=on')
+ else:
+ _oc_cmd = 'oc'
+ if node.host.containerized:
+@@ -223,6 +223,6 @@ class ocp(Cluster):
+
+ def set_node_options(self, node):
+ # don't attempt OC API collections on non-primary nodes
+- node.plugin_options.append('openshift.no-oc=on')
++ node.plugopts.append('openshift.no-oc=on')
+
+ # vim: set et ts=4 sw=4 :
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 8a9dbd7a..ab7f23cc 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -714,7 +714,7 @@ class SosNode():
+ elif line.startswith("The checksum is: "):
+ checksum = line.split()[3]
+
+- if checksum is not None:
++ if checksum:
+ self.manifest.add_field('checksum', checksum)
+ if len(checksum) == 32:
+ self.manifest.add_field('checksum_type', 'md5')
+@@ -722,6 +722,8 @@ class SosNode():
+ self.manifest.add_field('checksum_type', 'sha256')
+ else:
+ self.manifest.add_field('checksum_type', 'unknown')
++ else:
++ self.manifest.add_field('checksum_type', 'unknown')
+ else:
+ err = self.determine_sos_error(res['status'], res['output'])
+ self.log_debug("Error running sos report. rc = %s msg = %s"
+--
+2.31.1
+
+From 38a0533de3dd2613eefcc4865a2916e225e3ceed Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Tue, 9 Nov 2021 19:34:25 +0100
+Subject: [PATCH] [presets] Optimise OCP preset for hundreds of network
+ namespaces
+
+Sos report on OCP having hundreds of namespaces timeouts in networking
+plugin, as it collects >10 commands for each namespace.
+
+Let use a balanced approach in:
+- increasing network.timeout
+- limiting namespaces to traverse
+- disabling ethtool per namespace
+
+to ensure sos report successfully finish in a reasonable time,
+collecting rasonable amount of data.
+
+Resolves: #2754
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/presets/redhat/__init__.py | 10 +++++++---
+ 1 file changed, 7 insertions(+), 3 deletions(-)
+
+diff --git a/sos/presets/redhat/__init__.py b/sos/presets/redhat/__init__.py
+index e6d63611..865c9b6b 100644
+--- a/sos/presets/redhat/__init__.py
++++ b/sos/presets/redhat/__init__.py
+@@ -29,11 +29,15 @@ RHEL_DESC = RHEL_RELEASE_STR
+
+ RHOSP = "rhosp"
+ RHOSP_DESC = "Red Hat OpenStack Platform"
++RHOSP_OPTS = SoSOptions(plugopts=[
++ 'process.lsof=off',
++ 'networking.ethtool_namespaces=False',
++ 'networking.namespaces=200'])
+
+ RHOCP = "ocp"
+ RHOCP_DESC = "OpenShift Container Platform by Red Hat"
+-RHOSP_OPTS = SoSOptions(plugopts=[
+- 'process.lsof=off',
++RHOCP_OPTS = SoSOptions(all_logs=True, verify=True, plugopts=[
++ 'networking.timeout=600',
+ 'networking.ethtool_namespaces=False',
+ 'networking.namespaces=200'])
+
+@@ -62,7 +66,7 @@ RHEL_PRESETS = {
+ RHEL: PresetDefaults(name=RHEL, desc=RHEL_DESC),
+ RHOSP: PresetDefaults(name=RHOSP, desc=RHOSP_DESC, opts=RHOSP_OPTS),
+ RHOCP: PresetDefaults(name=RHOCP, desc=RHOCP_DESC, note=NOTE_SIZE_TIME,
+- opts=_opts_all_logs_verify),
++ opts=RHOCP_OPTS),
+ RH_CFME: PresetDefaults(name=RH_CFME, desc=RH_CFME_DESC, note=NOTE_TIME,
+ opts=_opts_verify),
+ RH_SATELLITE: PresetDefaults(name=RH_SATELLITE, desc=RH_SATELLITE_DESC,
+--
+2.31.1
+
+From 97b93c7f8755d04bdeb4f93759c20dcb787f2046 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 2 Nov 2021 11:34:13 -0400
+Subject: [PATCH] [Plugin] Rework get_container_logs to be more useful
+
+`get_container_logs()` is now `add_container_logs()` to align it better
+with our more common `add_*` methods for plugin collections.
+
+Additionally, it has been extended to accept either a single string or a
+list of strings like the other methods, and plugin authors may now
+specify either specific container names or regexes.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/__init__.py | 22 +++++++++++++++++-----
+ sos/report/plugins/rabbitmq.py | 2 +-
+ 2 files changed, 18 insertions(+), 6 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 08eee118..4b0e4fd5 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -2366,20 +2366,32 @@ class Plugin():
+ return _runtime.volumes
+ return []
+
+- def get_container_logs(self, container, **kwargs):
+- """Helper to get the ``logs`` output for a given container
++ def add_container_logs(self, containers, get_all=False, **kwargs):
++ """Helper to get the ``logs`` output for a given container or list
++ of container names and/or regexes.
+
+ Supports passthru of add_cmd_output() options
+
+- :param container: The name of the container to retrieve logs from
+- :type container: ``str``
++ :param containers: The name of the container to retrieve logs from,
++ may be a single name or a regex
++ :type containers: ``str`` or ``list` of strs
++
++ :param get_all: Should non-running containers also be queried?
++ Default: False
++ :type get_all: ``bool``
+
+ :param kwargs: Any kwargs supported by ``add_cmd_output()`` are
+ supported here
+ """
+ _runtime = self._get_container_runtime()
+ if _runtime is not None:
+- self.add_cmd_output(_runtime.get_logs_command(container), **kwargs)
++ if isinstance(containers, str):
++ containers = [containers]
++ for container in containers:
++ _cons = self.get_all_containers_by_regex(container, get_all)
++ for _con in _cons:
++ cmd = _runtime.get_logs_command(_con[1])
++ self.add_cmd_output(cmd, **kwargs)
+
+ def fmt_container_cmd(self, container, cmd, quotecmd=False):
+ """Format a command to be executed by the loaded ``ContainerRuntime``
+diff --git a/sos/report/plugins/rabbitmq.py b/sos/report/plugins/rabbitmq.py
+index e84b52da..1bfa741f 100644
+--- a/sos/report/plugins/rabbitmq.py
++++ b/sos/report/plugins/rabbitmq.py
+@@ -32,7 +32,7 @@ class RabbitMQ(Plugin, IndependentPlugin):
+
+ if in_container:
+ for container in container_names:
+- self.get_container_logs(container)
++ self.add_container_logs(container)
+ self.add_cmd_output(
+ self.fmt_container_cmd(container, 'rabbitmqctl report'),
+ foreground=True
+--
+2.31.1
+
+From 8bf602108f75db10e449eff5e2266c6466504086 Mon Sep 17 00:00:00 2001
+From: Nadia Pinaeva <npinaeva@redhat.com>
+Date: Thu, 2 Dec 2021 16:30:44 +0100
+Subject: [PATCH] [clusters:ocp] fix get_nodes function
+
+Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
+---
+ sos/collector/clusters/ocp.py | 8 ++++----
+ 1 file changed, 4 insertions(+), 4 deletions(-)
+
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index 22a7289a..2ce4e977 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -150,13 +150,13 @@ class ocp(Cluster):
+ "role option with '-c ocp.role=role1:role2'")
+ roles = [r for r in self.get_option('role').split(':')]
+ self.node_dict = self._build_dict(res['output'].splitlines())
+- for node in self.node_dict:
++ for node_name, node in self.node_dict.items():
+ if roles:
+ for role in roles:
+- if role in node:
+- nodes.append(node)
++ if role == node['roles']:
++ nodes.append(node_name)
+ else:
+- nodes.append(node)
++ nodes.append(node_name)
+ else:
+ msg = "'oc' command failed"
+ if 'Missing or incomplete' in res['output']:
+--
+2.31.1
+
+From 5d80ac6dc67e12ef00903436c088a1694f9a7dd7 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 1 Dec 2021 14:13:16 -0500
+Subject: [PATCH] [collect] Catch command not found exceptions from pexpect
+
+When running a command that does not exist on the system, catch the
+resulting pexpect exception and return the proper status code rather
+than allowing an untrapped exception.
+
+Closes: #2768
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/transports/__init__.py | 6 +++++-
+ 1 file changed, 5 insertions(+), 1 deletion(-)
+
+diff --git a/sos/collector/transports/__init__.py b/sos/collector/transports/__init__.py
+index 7bffee62..33f2f66d 100644
+--- a/sos/collector/transports/__init__.py
++++ b/sos/collector/transports/__init__.py
+@@ -225,7 +225,11 @@ class RemoteTransport():
+ if not env:
+ env = None
+
+- result = pexpect.spawn(cmd, encoding='utf-8', env=env)
++ try:
++ result = pexpect.spawn(cmd, encoding='utf-8', env=env)
++ except pexpect.exceptions.ExceptionPexpect as err:
++ self.log_debug(err.value)
++ return {'status': 127, 'output': ''}
+
+ _expects = [pexpect.EOF, pexpect.TIMEOUT]
+ if need_root and self.opts.ssh_user != 'root':
+--
+2.31.1
+
+From decb5d26c165e664fa879a669f2d80165181f0e1 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 2 Dec 2021 14:02:17 -0500
+Subject: [PATCH] [report,collect] Add option to control default container
+ runtime
+
+Adds a new `--container-runtime` option that allows users to control
+what default container runtime is used by plugins for container based
+collections, effectively overriding policy defaults.
+
+If no runtimes are active, this option is effectively ignored. If
+however runtimes are active, but the requested one is not, raise an
+exception to abort collection with an appropriate message to the user.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ man/en/sos-collect.1 | 6 ++++++
+ man/en/sos-report.1 | 19 +++++++++++++++++++
+ sos/collector/__init__.py | 4 ++++
+ sos/collector/sosnode.py | 6 ++++++
+ sos/report/__init__.py | 36 ++++++++++++++++++++++++++++++++++++
+ 5 files changed, 71 insertions(+)
+
+diff --git a/man/en/sos-collect.1 b/man/en/sos-collect.1
+index a1f6c10e..9b0a5d7b 100644
+--- a/man/en/sos-collect.1
++++ b/man/en/sos-collect.1
+@@ -11,6 +11,7 @@ sos collect \- Collect sosreports from multiple (cluster) nodes
+ [\-\-chroot CHROOT]
+ [\-\-case\-id CASE_ID]
+ [\-\-cluster\-type CLUSTER_TYPE]
++ [\-\-container\-runtime RUNTIME]
+ [\-e ENABLE_PLUGINS]
+ [--encrypt-key KEY]\fR
+ [--encrypt-pass PASS]\fR
+@@ -113,6 +114,11 @@ Example: \fBsos collect --cluster-type=kubernetes\fR will force the kubernetes p
+ to be run, and thus set sosreport options and attempt to determine a list of nodes using
+ that profile.
+ .TP
++\fB\-\-container\-runtime\fR RUNTIME
++\fB sos report\fR option. Using this with \fBcollect\fR will pass this option thru
++to nodes with sos version 4.3 or later. This option controls the default container
++runtime plugins will use for collections. See \fBman sos-report\fR.
++.TP
+ \fB\-e\fR ENABLE_PLUGINS, \fB\-\-enable\-plugins\fR ENABLE_PLUGINS
+ Sosreport option. Use this to enable a plugin that would otherwise not be run.
+
+diff --git a/man/en/sos-report.1 b/man/en/sos-report.1
+index e8efc8f8..464a77e5 100644
+--- a/man/en/sos-report.1
++++ b/man/en/sos-report.1
+@@ -19,6 +19,7 @@ sos report \- Collect and package diagnostic and support data
+ [--plugin-timeout TIMEOUT]\fR
+ [--cmd-timeout TIMEOUT]\fR
+ [--namespaces NAMESPACES]\fR
++ [--container-runtime RUNTIME]\fR
+ [-s|--sysroot SYSROOT]\fR
+ [-c|--chroot {auto|always|never}\fR
+ [--tmp-dir directory]\fR
+@@ -299,6 +300,24 @@ Use '0' (default) for no limit - all namespaces will be used for collections.
+
+ Note that specific plugins may provide a similar `namespaces` plugin option. If
+ the plugin option is used, it will override this option.
++.TP
++.B \--container-runtime RUNTIME
++Force the use of the specified RUNTIME as the default runtime that plugins will
++use to collect data from and about containers and container images. By default,
++the setting of \fBauto\fR results in the local policy determining what runtime
++will be the default runtime (in configurations where multiple runtimes are installed
++and active).
++
++If no container runtimes are active, this option is ignored. If there are runtimes
++active, but not one with a name matching RUNTIME, sos will abort.
++
++Setting this to \fBnone\fR, \fBoff\fR, or \fBdisabled\fR will cause plugins to
++\fBNOT\fR leverage any active runtimes for collections. Note that if disabled, plugins
++specifically for runtimes (e.g. the podman or docker plugins) will still collect
++general data about the runtime, but will not inspect existing containers or images.
++
++Default: 'auto' (policy determined)
++.TP
+ .B \--case-id NUMBER
+ Specify a case identifier to associate with the archive.
+ Identifiers may include alphanumeric characters, commas and periods ('.').
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index 42a7731d..3ad703d3 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -55,6 +55,7 @@ class SoSCollector(SoSComponent):
+ 'clean': False,
+ 'cluster_options': [],
+ 'cluster_type': None,
++ 'container_runtime': 'auto',
+ 'domains': [],
+ 'enable_plugins': [],
+ 'encrypt_key': '',
+@@ -268,6 +269,9 @@ class SoSCollector(SoSComponent):
+ sos_grp.add_argument('--chroot', default='',
+ choices=['auto', 'always', 'never'],
+ help="chroot executed commands to SYSROOT")
++ sos_grp.add_argument("--container-runtime", default="auto",
++ help="Default container runtime to use for "
++ "collections. 'auto' for policy control.")
+ sos_grp.add_argument('-e', '--enable-plugins', action="extend",
+ help='Enable specific plugins for sosreport')
+ sos_grp.add_argument('-k', '--plugin-option', '--plugopts',
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index ab7f23cc..f5957e17 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -586,6 +586,12 @@ class SosNode():
+ sos_opts.append('--cmd-timeout=%s'
+ % quote(str(self.opts.cmd_timeout)))
+
++ if self.check_sos_version('4.3'):
++ if self.opts.container_runtime != 'auto':
++ sos_opts.append(
++ "--container-runtime=%s" % self.opts.container_runtime
++ )
++
+ self.update_cmd_from_cluster()
+
+ sos_cmd = sos_cmd.replace(
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index a6c72778..0daad82f 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -82,6 +82,7 @@ class SoSReport(SoSComponent):
+ 'case_id': '',
+ 'chroot': 'auto',
+ 'clean': False,
++ 'container_runtime': 'auto',
+ 'keep_binary_files': False,
+ 'desc': '',
+ 'domains': [],
+@@ -187,6 +188,7 @@ class SoSReport(SoSComponent):
+ self.tempfile_util.clean()
+ self._exit(1)
+
++ self._check_container_runtime()
+ self._get_hardware_devices()
+ self._get_namespaces()
+
+@@ -218,6 +220,9 @@ class SoSReport(SoSComponent):
+ dest="chroot", default='auto',
+ help="chroot executed commands to SYSROOT "
+ "[auto, always, never] (default=auto)")
++ report_grp.add_argument("--container-runtime", default="auto",
++ help="Default container runtime to use for "
++ "collections. 'auto' for policy control.")
+ report_grp.add_argument("--desc", "--description", type=str,
+ action="store", default="",
+ help="Description for a new preset",)
+@@ -373,6 +378,37 @@ class SoSReport(SoSComponent):
+ }
+ # TODO: enumerate network devices, preferably with devtype info
+
++ def _check_container_runtime(self):
++ """Check the loaded container runtimes, and the policy default runtime
++ (if set), against any requested --container-runtime value. This can be
++ useful for systems that have multiple runtimes, such as RHCOS, but do
++ not have a clearly defined 'default' (or one that is determined based
++ entirely on configuration).
++ """
++ if self.opts.container_runtime != 'auto':
++ crun = self.opts.container_runtime.lower()
++ if crun in ['none', 'off', 'diabled']:
++ self.policy.runtimes = {}
++ self.soslog.info(
++ "Disabled all container runtimes per user option."
++ )
++ elif not self.policy.runtimes:
++ msg = ("WARNING: No container runtimes are active, ignoring "
++ "option to set default runtime to '%s'\n" % crun)
++ self.soslog.warn(msg)
++ elif crun not in self.policy.runtimes.keys():
++ valid = ', '.join(p for p in self.policy.runtimes.keys()
++ if p != 'default')
++ raise Exception("Cannot use container runtime '%s': no such "
++ "runtime detected. Available runtimes: %s"
++ % (crun, valid))
++ else:
++ self.policy.runtimes['default'] = self.policy.runtimes[crun]
++ self.soslog.info(
++ "Set default container runtime to '%s'"
++ % self.policy.runtimes['default'].name
++ )
++
+ def get_fibre_devs(self):
+ """Enumerate a list of fibrechannel devices on this system so that
+ plugins can iterate over them
+--
+2.31.1
+
+From 9d4b5af39d76ac99afa40d004fe9888633218356 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Fri, 3 Dec 2021 13:37:09 -0500
+Subject: [PATCH 1/2] [Plugin] Add container parameter for add_cmd_output()
+
+Adds a new `container` parameter for `Plugin.add_cmd_output()`, which if
+set will format all commands passed to that call for execution in the
+specified container.
+
+`Plugin.fmt_container_cmd()` is called for this purpose, and has been
+modified so that if the given container does not exist, an empty string
+is returned instead, thus preventing execution on the host.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/__init__.py | 16 ++++++++++++++--
+ 1 file changed, 14 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index e180ae17..3ff7c191 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -1707,7 +1707,7 @@ class Plugin():
+ chroot=True, runat=None, env=None, binary=False,
+ sizelimit=None, pred=None, subdir=None,
+ changes=False, foreground=False, tags=[],
+- priority=10, cmd_as_tag=False):
++ priority=10, cmd_as_tag=False, container=None):
+ """Run a program or a list of programs and collect the output
+
+ Output will be limited to `sizelimit`, collecting the last X amount
+@@ -1772,6 +1772,10 @@ class Plugin():
+ :param cmd_as_tag: Should the command string be automatically formatted
+ to a tag?
+ :type cmd_as_tag: ``bool``
++
++ :param container: Run the specified `cmds` inside a container with this
++ ID or name
++ :type container: ``str``
+ """
+ if isinstance(cmds, str):
+ cmds = [cmds]
+@@ -1782,6 +1786,14 @@ class Plugin():
+ if pred is None:
+ pred = self.get_predicate(cmd=True)
+ for cmd in cmds:
++ if container:
++ ocmd = cmd
++ cmd = self.fmt_container_cmd(container, cmd)
++ if not cmd:
++ self._log_debug("Skipping command '%s' as the requested "
++ "container '%s' does not exist."
++ % (ocmd, container))
++ continue
+ self._add_cmd_output(cmd=cmd, suggest_filename=suggest_filename,
+ root_symlink=root_symlink, timeout=timeout,
+ stderr=stderr, chroot=chroot, runat=runat,
+@@ -2420,7 +2432,7 @@ class Plugin():
+ if self.container_exists(container):
+ _runtime = self._get_container_runtime()
+ return _runtime.fmt_container_cmd(container, cmd, quotecmd)
+- return cmd
++ return ''
+
+ def is_module_loaded(self, module_name):
+ """Determine whether specified module is loaded or not
+--
+2.31.1
+
+
+From 874d2adfbff9e51dc902669af3c4a5083dbc19b1 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Fri, 3 Dec 2021 14:49:43 -0500
+Subject: [PATCH 2/2] [plugins] Update existing plugins to use a_c_o container
+ parameter
+
+Updates plugins currently calling `fmt_container_cmd()` in their
+`add_cmd_output()` calls to instead use the new `container` parameter
+and rely on the automatic formatting.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/opencontrail.py | 3 +--
+ sos/report/plugins/openstack_database.py | 20 ++++++--------------
+ sos/report/plugins/openstack_designate.py | 6 ++----
+ sos/report/plugins/openstack_ironic.py | 3 +--
+ sos/report/plugins/ovn_central.py | 7 +++----
+ sos/report/plugins/rabbitmq.py | 11 ++++++-----
+ 9 files changed, 47 insertions(+), 69 deletions(-)
+
+diff --git a/sos/report/plugins/opencontrail.py b/sos/report/plugins/opencontrail.py
+index b368bffe..76c03e21 100644
+--- a/sos/report/plugins/opencontrail.py
++++ b/sos/report/plugins/opencontrail.py
+@@ -25,8 +25,7 @@ class OpenContrail(Plugin, IndependentPlugin):
+ cnames = self.get_containers(get_all=True)
+ cnames = [c[1] for c in cnames if 'opencontrail' in c[1]]
+ for cntr in cnames:
+- _cmd = self.fmt_container_cmd(cntr, 'contrail-status')
+- self.add_cmd_output(_cmd)
++ self.add_cmd_output('contrail-status', container=cntr)
+ else:
+ self.add_cmd_output("contrail-status")
+
+diff --git a/sos/report/plugins/openstack_database.py b/sos/report/plugins/openstack_database.py
+index 1e98fabf..e9f84cf8 100644
+--- a/sos/report/plugins/openstack_database.py
++++ b/sos/report/plugins/openstack_database.py
+@@ -37,36 +37,28 @@ class OpenStackDatabase(Plugin):
+ ]
+
+ def setup(self):
+-
+- in_container = False
+ # determine if we're running databases on the host or in a container
+ _db_containers = [
+ 'galera-bundle-.*', # overcloud
+ 'mysql' # undercloud
+ ]
+
++ cname = None
+ for container in _db_containers:
+ cname = self.get_container_by_name(container)
+- if cname is not None:
+- in_container = True
++ if cname:
+ break
+
+- if in_container:
+- fname = "clustercheck_%s" % cname
+- cmd = self.fmt_container_cmd(cname, 'clustercheck')
+- self.add_cmd_output(cmd, timeout=15, suggest_filename=fname)
+- else:
+- self.add_cmd_output('clustercheck', timeout=15)
++ fname = "clustercheck_%s" % cname if cname else None
++ self.add_cmd_output('clustercheck', container=cname, timeout=15,
++ suggest_filename=fname)
+
+ if self.get_option('dump') or self.get_option('dumpall'):
+ db_dump = self.get_mysql_db_string(container=cname)
+ db_cmd = "mysqldump --opt %s" % db_dump
+
+- if in_container:
+- db_cmd = self.fmt_container_cmd(cname, db_cmd)
+-
+ self.add_cmd_output(db_cmd, suggest_filename='mysql_dump.sql',
+- sizelimit=0)
++ sizelimit=0, container=cname)
+
+ def get_mysql_db_string(self, container=None):
+
+diff --git a/sos/report/plugins/openstack_designate.py b/sos/report/plugins/openstack_designate.py
+index 0ae991b0..a2ea37ab 100644
+--- a/sos/report/plugins/openstack_designate.py
++++ b/sos/report/plugins/openstack_designate.py
+@@ -20,12 +20,10 @@ class OpenStackDesignate(Plugin):
+
+ def setup(self):
+ # collect current pool config
+- pools_cmd = self.fmt_container_cmd(
+- self.get_container_by_name(".*designate_central"),
+- "designate-manage pool generate_file --file /dev/stdout")
+
+ self.add_cmd_output(
+- pools_cmd,
++ "designate-manage pool generate_file --file /dev/stdout",
++ container=self.get_container_by_name(".*designate_central"),
+ suggest_filename="openstack_designate_current_pools.yaml"
+ )
+
+diff --git a/sos/report/plugins/openstack_ironic.py b/sos/report/plugins/openstack_ironic.py
+index c36fb6b6..49beb2d9 100644
+--- a/sos/report/plugins/openstack_ironic.py
++++ b/sos/report/plugins/openstack_ironic.py
+@@ -80,8 +80,7 @@ class OpenStackIronic(Plugin):
+ 'ironic_pxe_tftp', 'ironic_neutron_agent',
+ 'ironic_conductor', 'ironic_api']:
+ if self.container_exists('.*' + container_name):
+- self.add_cmd_output(self.fmt_container_cmd(container_name,
+- 'rpm -qa'))
++ self.add_cmd_output('rpm -qa', container=container_name)
+
+ else:
+ self.conf_list = [
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index 914eda60..ddbf288d 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -123,11 +123,10 @@ class OVNCentral(Plugin):
+
+ # If OVN is containerized, we need to run the above commands inside
+ # the container.
+- cmds = [
+- self.fmt_container_cmd(self._container_name, cmd) for cmd in cmds
+- ]
+
+- self.add_cmd_output(cmds, foreground=True)
++ self.add_cmd_output(
++ cmds, foreground=True, container=self._container_name
++ )
+
+ self.add_copy_spec("/etc/sysconfig/ovn-northd")
+
+diff --git a/sos/report/plugins/rabbitmq.py b/sos/report/plugins/rabbitmq.py
+index 1bfa741f..607802e4 100644
+--- a/sos/report/plugins/rabbitmq.py
++++ b/sos/report/plugins/rabbitmq.py
+@@ -34,14 +34,15 @@ class RabbitMQ(Plugin, IndependentPlugin):
+ for container in container_names:
+ self.add_container_logs(container)
+ self.add_cmd_output(
+- self.fmt_container_cmd(container, 'rabbitmqctl report'),
++ 'rabbitmqctl report',
++ container=container,
+ foreground=True
+ )
+ self.add_cmd_output(
+- self.fmt_container_cmd(
+- container, "rabbitmqctl eval "
+- "'rabbit_diagnostics:maybe_stuck().'"),
+- foreground=True, timeout=10
++ "rabbitmqctl eval 'rabbit_diagnostics:maybe_stuck().'",
++ container=container,
++ foreground=True,
++ timeout=10
+ )
+ else:
+ self.add_cmd_output("rabbitmqctl report")
+--
+2.31.1
+
+From faa15754f82e9841cd624afe18dc2198644decdf Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Wed, 8 Dec 2021 13:51:20 -0500
+Subject: [PATCH] [Policy,collect] Prevent remote node policies from setting
+ local PATH
+
+This commit fixes an issue where policies loaded for remote nodes when
+using `sos collect` would override the PATH setting for the local
+policy, which in turn could prevent successful execution of cluster
+profile operations.
+
+Related: #2777
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/policies/__init__.py | 8 +++++---
+ sos/policies/distros/__init__.py | 6 ++++--
+ sos/policies/distros/debian.py | 3 ++-
+ sos/policies/distros/redhat.py | 6 ++++--
+ sos/policies/distros/suse.py | 3 ++-
+ 5 files changed, 17 insertions(+), 9 deletions(-)
+
+diff --git a/sos/policies/__init__.py b/sos/policies/__init__.py
+index ef9188de..826d03a1 100644
+--- a/sos/policies/__init__.py
++++ b/sos/policies/__init__.py
+@@ -45,7 +45,7 @@ def load(cache={}, sysroot=None, init=None, probe_runtime=True,
+ return cache['policy']
+
+
+-class Policy(object):
++class Policy():
+ """Policies represent distributions that sos supports, and define the way
+ in which sos behaves on those distributions. A policy should define at
+ minimum a way to identify the distribution, and a package manager to allow
+@@ -111,7 +111,7 @@ any third party.
+ presets_path = PRESETS_PATH
+ _in_container = False
+
+- def __init__(self, sysroot=None, probe_runtime=True):
++ def __init__(self, sysroot=None, probe_runtime=True, remote_exec=None):
+ """Subclasses that choose to override this initializer should call
+ super() to ensure that they get the required platform bits attached.
+ super(SubClass, self).__init__(). Policies that require runtime
+@@ -122,7 +122,9 @@ any third party.
+ self.probe_runtime = probe_runtime
+ self.package_manager = PackageManager()
+ self.valid_subclasses = [IndependentPlugin]
+- self.set_exec_path()
++ self.remote_exec = remote_exec
++ if not self.remote_exec:
++ self.set_exec_path()
+ self.sysroot = sysroot
+ self.register_presets(GENERIC_PRESETS)
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index c69fc1e7..9c91a918 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -68,9 +68,11 @@ class LinuxPolicy(Policy):
+ container_version_command = None
+ container_authfile = None
+
+- def __init__(self, sysroot=None, init=None, probe_runtime=True):
++ def __init__(self, sysroot=None, init=None, probe_runtime=True,
++ remote_exec=None):
+ super(LinuxPolicy, self).__init__(sysroot=sysroot,
+- probe_runtime=probe_runtime)
++ probe_runtime=probe_runtime,
++ remote_exec=remote_exec)
+
+ if sysroot:
+ self.sysroot = sysroot
+diff --git a/sos/policies/distros/debian.py b/sos/policies/distros/debian.py
+index 639fd5eb..41f09428 100644
+--- a/sos/policies/distros/debian.py
++++ b/sos/policies/distros/debian.py
+@@ -26,7 +26,8 @@ class DebianPolicy(LinuxPolicy):
+ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ remote_exec=None):
+ super(DebianPolicy, self).__init__(sysroot=sysroot, init=init,
+- probe_runtime=probe_runtime)
++ probe_runtime=probe_runtime,
++ remote_exec=remote_exec)
+ self.package_manager = DpkgPackageManager(chroot=self.sysroot,
+ remote_exec=remote_exec)
+ self.valid_subclasses += [DebianPlugin]
+diff --git a/sos/policies/distros/redhat.py b/sos/policies/distros/redhat.py
+index 4b14abaf..eb75e15b 100644
+--- a/sos/policies/distros/redhat.py
++++ b/sos/policies/distros/redhat.py
+@@ -53,7 +53,8 @@ class RedHatPolicy(LinuxPolicy):
+ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ remote_exec=None):
+ super(RedHatPolicy, self).__init__(sysroot=sysroot, init=init,
+- probe_runtime=probe_runtime)
++ probe_runtime=probe_runtime,
++ remote_exec=remote_exec)
+ self.usrmove = False
+
+ self.package_manager = RpmPackageManager(chroot=self.sysroot,
+@@ -76,7 +77,8 @@ class RedHatPolicy(LinuxPolicy):
+ self.PATH = "/sbin:/bin:/usr/sbin:/usr/bin:/root/bin"
+ self.PATH += os.pathsep + "/usr/local/bin"
+ self.PATH += os.pathsep + "/usr/local/sbin"
+- self.set_exec_path()
++ if not self.remote_exec:
++ self.set_exec_path()
+ self.load_presets()
+
+ @classmethod
+diff --git a/sos/policies/distros/suse.py b/sos/policies/distros/suse.py
+index 1c1feff5..b9d4a3b1 100644
+--- a/sos/policies/distros/suse.py
++++ b/sos/policies/distros/suse.py
+@@ -25,7 +25,8 @@ class SuSEPolicy(LinuxPolicy):
+ def __init__(self, sysroot=None, init=None, probe_runtime=True,
+ remote_exec=None):
+ super(SuSEPolicy, self).__init__(sysroot=sysroot, init=init,
+- probe_runtime=probe_runtime)
++ probe_runtime=probe_runtime,
++ remote_exec=remote_exec)
+ self.valid_subclasses += [SuSEPlugin, RedHatPlugin]
+
+ self.usrmove = False
+--
+2.31.1
+
+From d4383fec5f8a80121aa4f5a37575b37988c51663 Mon Sep 17 00:00:00 2001
+From: Nadia Pinaeva <npinaeva@redhat.com>
+Date: Wed, 1 Dec 2021 12:23:34 +0100
+Subject: [PATCH] Add crio runtime and openshift_ovn plugin openshift_ovn
+ plugin collects logs from crio containers Fix get_container_by_name function
+ returning container_id and not name
+
+Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
+---
+ sos/policies/distros/__init__.py | 4 +-
+ sos/policies/runtimes/__init__.py | 2 +-
+ sos/policies/runtimes/crio.py | 79 +++++++++++++++++++++++++++++
+ sos/report/plugins/openshift_ovn.py | 41 +++++++++++++++
+ 4 files changed, 124 insertions(+), 2 deletions(-)
+ create mode 100644 sos/policies/runtimes/crio.py
+ create mode 100644 sos/report/plugins/openshift_ovn.py
+
+diff --git a/sos/policies/distros/__init__.py b/sos/policies/distros/__init__.py
+index 9c91a918..7acc7e49 100644
+--- a/sos/policies/distros/__init__.py
++++ b/sos/policies/distros/__init__.py
+@@ -17,6 +17,7 @@ from sos import _sos as _
+ from sos.policies import Policy
+ from sos.policies.init_systems import InitSystem
+ from sos.policies.init_systems.systemd import SystemdInit
++from sos.policies.runtimes.crio import CrioContainerRuntime
+ from sos.policies.runtimes.podman import PodmanContainerRuntime
+ from sos.policies.runtimes.docker import DockerContainerRuntime
+
+@@ -92,7 +93,8 @@ class LinuxPolicy(Policy):
+ if self.probe_runtime:
+ _crun = [
+ PodmanContainerRuntime(policy=self),
+- DockerContainerRuntime(policy=self)
++ DockerContainerRuntime(policy=self),
++ CrioContainerRuntime(policy=self)
+ ]
+ for runtime in _crun:
+ if runtime.check_is_active():
+diff --git a/sos/policies/runtimes/__init__.py b/sos/policies/runtimes/__init__.py
+index 2e60ad23..4e9a45c1 100644
+--- a/sos/policies/runtimes/__init__.py
++++ b/sos/policies/runtimes/__init__.py
+@@ -100,7 +100,7 @@ class ContainerRuntime():
+ return None
+ for c in self.containers:
+ if re.match(name, c[1]):
+- return c[1]
++ return c[0]
+ return None
+
+ def get_images(self):
+diff --git a/sos/policies/runtimes/crio.py b/sos/policies/runtimes/crio.py
+new file mode 100644
+index 00000000..980c3ea1
+--- /dev/null
++++ b/sos/policies/runtimes/crio.py
+@@ -0,0 +1,79 @@
++# Copyright (C) 2021 Red Hat, Inc., Nadia Pinaeva <npinaeva@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++from sos.policies.runtimes import ContainerRuntime
++from sos.utilities import sos_get_command_output
++from pipes import quote
++
++
++class CrioContainerRuntime(ContainerRuntime):
++ """Runtime class to use for systems running crio"""
++
++ name = 'crio'
++ binary = 'crictl'
++
++ def get_containers(self, get_all=False):
++ """Get a list of containers present on the system.
++
++ :param get_all: If set, include stopped containers as well
++ :type get_all: ``bool``
++ """
++ containers = []
++ _cmd = "%s ps %s" % (self.binary, '-a' if get_all else '')
++ if self.active:
++ out = sos_get_command_output(_cmd, chroot=self.policy.sysroot)
++ if out['status'] == 0:
++ for ent in out['output'].splitlines()[1:]:
++ ent = ent.split()
++ # takes the form (container_id, container_name)
++ containers.append((ent[0], ent[-3]))
++ return containers
++
++ def get_images(self):
++ """Get a list of images present on the system
++
++ :returns: A list of 2-tuples containing (image_name, image_id)
++ :rtype: ``list``
++ """
++ images = []
++ if self.active:
++ out = sos_get_command_output("%s images" % self.binary,
++ chroot=self.policy.sysroot)
++ if out['status'] == 0:
++ for ent in out['output'].splitlines():
++ ent = ent.split()
++ # takes the form (image_name, image_id)
++ images.append((ent[0] + ':' + ent[1], ent[2]))
++ return images
++
++ def fmt_container_cmd(self, container, cmd, quotecmd):
++ """Format a command to run inside a container using the runtime
++
++ :param container: The name or ID of the container in which to run
++ :type container: ``str``
++
++ :param cmd: The command to run inside `container`
++ :type cmd: ``str``
++
++ :param quotecmd: Whether the cmd should be quoted.
++ :type quotecmd: ``bool``
++
++ :returns: Formatted string to run `cmd` inside `container`
++ :rtype: ``str``
++ """
++ if quotecmd:
++ quoted_cmd = quote(cmd)
++ else:
++ quoted_cmd = cmd
++ container_id = self.get_container_by_name(container)
++ return "%s %s %s" % (self.run_cmd, container_id,
++ quoted_cmd) if container_id is not None else ''
++
++# vim: set et ts=4 sw=4 :
+diff --git a/sos/report/plugins/openshift_ovn.py b/sos/report/plugins/openshift_ovn.py
+new file mode 100644
+index 00000000..168f1dd3
+--- /dev/null
++++ b/sos/report/plugins/openshift_ovn.py
+@@ -0,0 +1,41 @@
++# Copyright (C) 2021 Nadia Pinaeva <npinaeva@redhat.com>
++
++# This file is part of the sos project: https://github.com/sosreport/sos
++#
++# This copyrighted material is made available to anyone wishing to use,
++# modify, copy, or redistribute it subject to the terms and conditions of
++# version 2 of the GNU General Public License.
++#
++# See the LICENSE file in the source distribution for further information.
++
++from sos.report.plugins import Plugin, RedHatPlugin
++
++
++class OpenshiftOVN(Plugin, RedHatPlugin):
++ """This plugin is used to collect OCP 4.x OVN logs.
++ """
++ short_desc = 'Openshift OVN'
++ plugin_name = "openshift_ovn"
++ containers = ('ovnkube-master', 'ovn-ipsec')
++ profiles = ('openshift',)
++
++ def setup(self):
++ self.add_copy_spec([
++ "/var/lib/ovn/etc/ovnnb_db.db",
++ "/var/lib/ovn/etc/ovnsb_db.db",
++ "/var/lib/openvswitch/etc/keys",
++ "/var/log/openvswitch/libreswan.log",
++ "/var/log/openvswitch/ovs-monitor-ipsec.log"
++ ])
++
++ self.add_cmd_output([
++ 'ovn-appctl -t /var/run/ovn/ovnnb_db.ctl ' +
++ 'cluster/status OVN_Northbound',
++ 'ovn-appctl -t /var/run/ovn/ovnsb_db.ctl ' +
++ 'cluster/status OVN_Southbound'],
++ container='ovnkube-master')
++ self.add_cmd_output([
++ 'ovs-appctl -t ovs-monitor-ipsec tunnels/show',
++ 'ipsec status',
++ 'certutil -L -d sql:/etc/ipsec.d'],
++ container='ovn-ipsec')
+--
+2.31.1
+
+From 17218ca17e49cb8491c688095b56503d041c1ae9 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 9 Dec 2021 15:07:23 -0500
+Subject: [PATCH 1/3] [ocp] Skip project setup whenever oc transport is not
+ used
+
+Fixes a corner case where we would still attempt to create a new project
+within the OCP cluster even if we weren't using the `oc` transport.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/clusters/ocp.py | 4 +++-
+ 1 file changed, 3 insertions(+), 1 deletion(-)
+
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index 2ce4e977..56f8cc47 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -123,7 +123,9 @@ class ocp(Cluster):
+ return nodes
+
+ def set_transport_type(self):
+- if is_executable('oc') or self.opts.transport == 'oc':
++ if self.opts.transport != 'auto':
++ return self.opts.transport
++ if is_executable('oc'):
+ return 'oc'
+ self.log_info("Local installation of 'oc' not found or is not "
+ "correctly configured. Will use ControlPersist.")
+--
+2.31.1
+
+
+From 9faabdc3df08516a91c1adb3326bbf43db155f71 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 9 Dec 2021 16:04:39 -0500
+Subject: [PATCH 2/3] [crio] Put inspect output in the containers subdir
+
+Given the environments where crio is run, having `crictl inspect` output
+in the main plugin directory can be a bit overwhelming. As such, put
+this output into a `containers` subdir, and nest container log output in
+a `containers/logs/` subdir.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/crio.py | 5 +++--
+ 1 file changed, 3 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/crio.py b/sos/report/plugins/crio.py
+index cb2c9796..56cf64a7 100644
+--- a/sos/report/plugins/crio.py
++++ b/sos/report/plugins/crio.py
+@@ -79,10 +79,11 @@ class CRIO(Plugin, RedHatPlugin, UbuntuPlugin):
+ pods = self._get_crio_list(pod_cmd)
+
+ for container in containers:
+- self.add_cmd_output("crictl inspect %s" % container)
++ self.add_cmd_output("crictl inspect %s" % container,
++ subdir="containers")
+ if self.get_option('logs'):
+ self.add_cmd_output("crictl logs -t %s" % container,
+- subdir="containers", priority=100)
++ subdir="containers/logs", priority=100)
+
+ for image in images:
+ self.add_cmd_output("crictl inspecti %s" % image, subdir="images")
+--
+2.31.1
+
+
+From 9118562c47fb521da3eeeed1a8746d45aaa769e7 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 9 Dec 2021 16:06:06 -0500
+Subject: [PATCH 3/3] [networking] Put namespaced commands into subdirs
+
+Where networking namespaces are used, there tend to be large numbers of
+namespaces used. This in turn results in sos running and collecting very
+large numbers of namespaced commands.
+
+To aid in consumability, place these collections under a subdir for the
+namespace under another "namespaces" subdir within the plugin directory.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/report/plugins/networking.py | 27 ++++++++++++---------------
+ 1 file changed, 12 insertions(+), 15 deletions(-)
+
+diff --git a/sos/report/plugins/networking.py b/sos/report/plugins/networking.py
+index 80e24abb..bcb5e6ae 100644
+--- a/sos/report/plugins/networking.py
++++ b/sos/report/plugins/networking.py
+@@ -198,6 +198,7 @@ class Networking(Plugin):
+ pred=SoSPredicate(self, cmd_outputs=co6))
+ else None)
+ for namespace in namespaces:
++ _subdir = "namespaces/%s" % namespace
+ ns_cmd_prefix = cmd_prefix + namespace + " "
+ self.add_cmd_output([
+ ns_cmd_prefix + "ip address show",
+@@ -213,29 +214,27 @@ class Networking(Plugin):
+ ns_cmd_prefix + "netstat -s",
+ ns_cmd_prefix + "netstat %s -agn" % self.ns_wide,
+ ns_cmd_prefix + "nstat -zas",
+- ], priority=50)
++ ], priority=50, subdir=_subdir)
+ self.add_cmd_output([ns_cmd_prefix + "iptables-save"],
+ pred=iptables_with_nft,
++ subdir=_subdir,
+ priority=50)
+ self.add_cmd_output([ns_cmd_prefix + "ip6tables-save"],
+ pred=ip6tables_with_nft,
++ subdir=_subdir,
+ priority=50)
+
+ ss_cmd = ns_cmd_prefix + "ss -peaonmi"
+ # --allow-system-changes is handled directly in predicate
+ # evaluation, so plugin code does not need to separately
+ # check for it
+- self.add_cmd_output(ss_cmd, pred=ss_pred)
+-
+- # Collect ethtool commands only when ethtool_namespaces
+- # is set to true.
+- if self.get_option("ethtool_namespaces"):
+- # Devices that exist in a namespace use less ethtool
+- # parameters. Run this per namespace.
+- for namespace in self.get_network_namespaces(
+- self.get_option("namespace_pattern"),
+- self.get_option("namespaces")):
+- ns_cmd_prefix = cmd_prefix + namespace + " "
++ self.add_cmd_output(ss_cmd, pred=ss_pred, subdir=_subdir)
++
++ # Collect ethtool commands only when ethtool_namespaces
++ # is set to true.
++ if self.get_option("ethtool_namespaces"):
++ # Devices that exist in a namespace use less ethtool
++ # parameters. Run this per namespace.
+ netns_netdev_list = self.exec_cmd(
+ ns_cmd_prefix + "ls -1 /sys/class/net/"
+ )
+@@ -250,9 +249,7 @@ class Networking(Plugin):
+ ns_cmd_prefix + "ethtool -i " + eth,
+ ns_cmd_prefix + "ethtool -k " + eth,
+ ns_cmd_prefix + "ethtool -S " + eth
+- ], priority=50)
+-
+- return
++ ], priority=50, subdir=_subdir)
+
+
+ class RedHatNetworking(Networking, RedHatPlugin):
+--
+2.31.1
+
+From 4bf5f9143c962c839c1d27217ba74127551a5c00 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Fri, 17 Dec 2021 11:10:15 -0500
+Subject: [PATCH] [transport] Detect retrieval failures and automatically retry
+
+If a paritcular attempt to retrieve a remote file fails, we should
+automatically retry that collection up to a certain point. This provides
+`sos collect` more resiliency for the collection of sos report archives.
+
+This change necessitates a change in how we handle the SoSNode flow for
+failed sos report retrievals, and as such contains minor fixes to
+transports to ensure that we do not incorrectly hit exceptions in error
+handling that were not previously possible with how we exited the
+SoSNode retrieval flow.
+
+Closes: #2777
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/__init__.py | 5 +++--
+ sos/collector/clusters/ocp.py | 1 +
+ sos/collector/sosnode.py | 17 ++++++++++-------
+ sos/collector/transports/__init__.py | 15 ++++++++++++++-
+ sos/collector/transports/local.py | 1 +
+ sos/collector/transports/oc.py | 3 ++-
+ 6 files changed, 31 insertions(+), 11 deletions(-)
+
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index b825d8fc..a25e794e 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -1221,8 +1221,9 @@ this utility or remote systems that it connects to.
+ def close_all_connections(self):
+ """Close all sessions for nodes"""
+ for client in self.client_list:
+- self.log_debug('Closing connection to %s' % client.address)
+- client.disconnect()
++ if client.connected:
++ self.log_debug('Closing connection to %s' % client.address)
++ client.disconnect()
+
+ def create_cluster_archive(self):
+ """Calls for creation of tar archive then cleans up the temporary
+diff --git a/sos/collector/clusters/ocp.py b/sos/collector/clusters/ocp.py
+index 56f8cc47..ae93ad58 100644
+--- a/sos/collector/clusters/ocp.py
++++ b/sos/collector/clusters/ocp.py
+@@ -92,6 +92,7 @@ class ocp(Cluster):
+ % ret['output'])
+ # don't leave the config on a non-existing project
+ self.exec_master_cmd("oc project default")
++ self.project = None
+ return True
+
+ def _build_dict(self, nodelist):
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 1341e39f..925f2790 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -751,12 +751,11 @@ class SosNode():
+ if self.file_exists(path):
+ self.log_info("Copying remote %s to local %s" %
+ (path, destdir))
+- self._transport.retrieve_file(path, dest)
++ return self._transport.retrieve_file(path, dest)
+ else:
+ self.log_debug("Attempting to copy remote file %s, but it "
+ "does not exist on filesystem" % path)
+ return False
+- return True
+ except Exception as err:
+ self.log_debug("Failed to retrieve %s: %s" % (path, err))
+ return False
+@@ -793,16 +792,20 @@ class SosNode():
+ except Exception:
+ self.log_error('Failed to make archive readable')
+ return False
+- self.soslog.info('Retrieving sos report from %s' % self.address)
++ self.log_info('Retrieving sos report from %s' % self.address)
+ self.ui_msg('Retrieving sos report...')
+- ret = self.retrieve_file(self.sos_path)
++ try:
++ ret = self.retrieve_file(self.sos_path)
++ except Exception as err:
++ self.log_error(err)
++ return False
+ if ret:
+ self.ui_msg('Successfully collected sos report')
+ self.file_list.append(self.sos_path.split('/')[-1])
++ return True
+ else:
+- self.log_error('Failed to retrieve sos report')
+- raise SystemExit
+- return True
++ self.ui_msg('Failed to retrieve sos report')
++ return False
+ else:
+ # sos sometimes fails but still returns a 0 exit code
+ if self.stderr.read():
+diff --git a/sos/collector/transports/__init__.py b/sos/collector/transports/__init__.py
+index 33f2f66d..dcdebdde 100644
+--- a/sos/collector/transports/__init__.py
++++ b/sos/collector/transports/__init__.py
+@@ -303,7 +303,20 @@ class RemoteTransport():
+ :returns: True if file was successfully copied from remote, or False
+ :rtype: ``bool``
+ """
+- return self._retrieve_file(fname, dest)
++ attempts = 0
++ try:
++ while attempts < 5:
++ attempts += 1
++ ret = self._retrieve_file(fname, dest)
++ if ret:
++ return True
++ self.log_info("File retrieval attempt %s failed" % attempts)
++ self.log_info("File retrieval failed after 5 attempts")
++ return False
++ except Exception as err:
++ self.log_error("Exception encountered during retrieval attempt %s "
++ "for %s: %s" % (attempts, fname, err))
++ raise err
+
+ def _retrieve_file(self, fname, dest):
+ raise NotImplementedError("Transport %s does not support file copying"
+diff --git a/sos/collector/transports/local.py b/sos/collector/transports/local.py
+index a4897f19..2996d524 100644
+--- a/sos/collector/transports/local.py
++++ b/sos/collector/transports/local.py
+@@ -35,6 +35,7 @@ class LocalTransport(RemoteTransport):
+ def _retrieve_file(self, fname, dest):
+ self.log_debug("Moving %s to %s" % (fname, dest))
+ shutil.copy(fname, dest)
++ return True
+
+ def _format_cmd_for_exec(self, cmd):
+ return cmd
+diff --git a/sos/collector/transports/oc.py b/sos/collector/transports/oc.py
+index de044ccb..720dd61d 100644
+--- a/sos/collector/transports/oc.py
++++ b/sos/collector/transports/oc.py
+@@ -202,7 +202,8 @@ class OCTransport(RemoteTransport):
+ env, False)
+
+ def _disconnect(self):
+- os.unlink(self.pod_tmp_conf)
++ if os.path.exists(self.pod_tmp_conf):
++ os.unlink(self.pod_tmp_conf)
+ removed = self.run_oc("delete pod %s" % self.pod_name)
+ if "deleted" not in removed['output']:
+ self.log_debug("Calling delete on pod '%s' failed: %s"
+--
+2.31.1
+
+From 304c9ef6c1015f1ebe1a8d569c3e16bada4d23f1 Mon Sep 17 00:00:00 2001
+From: Nadia Pinaeva <npinaeva@redhat.com>
+Date: Tue, 4 Jan 2022 16:37:09 +0100
+Subject: [PATCH] Add cluster cleanup for all exit() calls
+
+Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
+---
+ sos/collector/__init__.py | 3 +--
+ 1 file changed, 1 insertion(+), 2 deletions(-)
+
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index a25e794e1..ffd63bc63 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -443,6 +443,7 @@ def add_parser_options(cls, parser):
+
+ def exit(self, msg, error=1):
+ """Used to safely terminate if sos-collector encounters an error"""
++ self.cluster.cleanup()
+ self.log_error(msg)
+ try:
+ self.close_all_connections()
+@@ -858,8 +858,9 @@ class SoSCollector(SoSComponent):
+ "CTRL-C to quit\n")
+ self.ui_log.info("")
+ except KeyboardInterrupt:
+- self.cluster.cleanup()
+ self.exit("Exiting on user cancel", 130)
++ except Exception as e:
++ self.exit(repr(e), 1)
+
+ def configure_sos_cmd(self):
+ """Configures the sosreport command that is run on the nodes"""
+@@ -1185,7 +1185,6 @@ def collect(self):
+ arc_name = self.create_cluster_archive()
+ else:
+ msg = 'No sosreports were collected, nothing to archive...'
+- self.cluster.cleanup()
+ self.exit(msg, 1)
+
+ if self.opts.upload and self.policy.get_upload_url():
+From 2c3a647817dfbac36be3768acf6026e91d1a6e8f Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 21 Dec 2021 14:20:19 -0500
+Subject: [PATCH] [options] Allow spaces in --keywords values in sos.conf
+
+The `--keywords` option supports spaces to allow for obfuscated phrases,
+not just words. This however breaks if a phrase is added to the config
+file *before* a run with the phrase in the cmdline option, due to the
+safeguards we have for all other values that do not support spaces.
+
+Add a check in our flow for updating options from the config file to not
+replace illegal spaces if we're checking the `keywords` option, for
+which spaces are legal.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/options.py | 5 ++++-
+ 1 file changed, 4 insertions(+), 1 deletion(-)
+
+diff --git a/sos/options.py b/sos/options.py
+index 7bea3ffc1..4846a5096 100644
+--- a/sos/options.py
++++ b/sos/options.py
+@@ -200,7 +200,10 @@ def _update_from_section(section, config):
+ odict[rename_opts[key]] = odict.pop(key)
+ # set the values according to the config file
+ for key, val in odict.items():
+- if isinstance(val, str):
++ # most option values do not tolerate spaces, special
++ # exception however for --keywords which we do want to
++ # support phrases, and thus spaces, for
++ if isinstance(val, str) and key != 'keywords':
+ val = val.replace(' ', '')
+ if key not in self.arg_defaults:
+ # read an option that is not loaded by the current
+From f912fc9e31b406a24b7a9c012e12cda920632051 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 10 Jan 2022 14:13:42 +0100
+Subject: [PATCH] [collect] Deal None sos_version properly
+
+In case collector cluster hits an error during init, sos_version
+is None what LooseVersion can't compare properly and raises exception
+
+'LooseVersion' object has no attribute 'version'
+
+Related: #2822
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/collector/sosnode.py | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 925f27909..7bbe0cd1f 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -382,7 +382,8 @@ def check_sos_version(self, ver):
+ given ver. This means that if the installed version is greater than
+ ver, this will still return True
+ """
+- return LooseVersion(self.sos_info['version']) >= ver
++ return self.sos_info['version'] is not None and \
++ LooseVersion(self.sos_info['version']) >= ver
+
+ def is_installed(self, pkg):
+ """Checks if a given package is installed on the node"""
+From 0c67e8ebaeef17dac3b5b9e42a59b4e673e4403b Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Mon, 10 Jan 2022 14:17:13 +0100
+Subject: [PATCH] [collector] Cleanup cluster only if defined
+
+In case cluster init fails, self.cluster = None and its cleanup
+must be skipped.
+
+Resolves: #2822
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/collector/__init__.py | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/sos/collector/__init__.py b/sos/collector/__init__.py
+index ffd63bc63..3e22bca3e 100644
+--- a/sos/collector/__init__.py
++++ b/sos/collector/__init__.py
+@@ -443,7 +443,8 @@ def add_parser_options(cls, parser):
+
+ def exit(self, msg, error=1):
+ """Used to safely terminate if sos-collector encounters an error"""
+- self.cluster.cleanup()
++ if self.cluster:
++ self.cluster.cleanup()
+ self.log_error(msg)
+ try:
+ self.close_all_connections()
+From ef27a6ee6737c23b3beda1437768a91679024697 Mon Sep 17 00:00:00 2001
+From: Nadia Pinaeva <npinaeva@redhat.com>
+Date: Fri, 3 Dec 2021 15:41:35 +0100
+Subject: [PATCH] Add journal logs for NetworkManager plugin
+
+Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
+---
+ sos/report/plugins/networkmanager.py | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/sos/report/plugins/networkmanager.py b/sos/report/plugins/networkmanager.py
+index 30f99a1140..3aca0c7460 100644
+--- a/sos/report/plugins/networkmanager.py
++++ b/sos/report/plugins/networkmanager.py
+@@ -25,6 +25,8 @@ def setup(self):
+ "/etc/NetworkManager/dispatcher.d"
+ ])
+
++ self.add_journal(units="NetworkManager")
++
+ # There are some incompatible changes in nmcli since
+ # the release of NetworkManager >= 0.9.9. In addition,
+ # NetworkManager >= 0.9.9 will use the long names of
+From 9eb60f0bb6ea36f9c1cf099c1fd20cf3938b4b26 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 17 Jan 2022 11:11:24 -0500
+Subject: [PATCH] [clean] Ignore empty items for obfuscation better
+
+This commit fixes a couple edge cases where an item empty (e.g. and
+empty string '') was not being properly ignored, which in turned caused
+failures in writing both obfuscations and replacement files.
+
+This should no longer be possible.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/mappings/__init__.py | 5 ++++-
+ sos/cleaner/mappings/username_map.py | 2 +-
+ sos/cleaner/parsers/username_parser.py | 2 +-
+ 3 files changed, 6 insertions(+), 3 deletions(-)
+
+diff --git a/sos/cleaner/mappings/__init__.py b/sos/cleaner/mappings/__init__.py
+index 5cf5c8b2d..48171a052 100644
+--- a/sos/cleaner/mappings/__init__.py
++++ b/sos/cleaner/mappings/__init__.py
+@@ -49,6 +49,8 @@ def add(self, item):
+ :param item: The plaintext object to obfuscate
+ """
+ with self.lock:
++ if not item:
++ return item
+ self.dataset[item] = self.sanitize_item(item)
+ return self.dataset[item]
+
+@@ -67,7 +69,8 @@ def get(self, item):
+ """Retrieve an item's obfuscated counterpart from the map. If the item
+ does not yet exist in the map, add it by generating one on the fly
+ """
+- if self.ignore_item(item) or self.item_in_dataset_values(item):
++ if (not item or self.ignore_item(item) or
++ self.item_in_dataset_values(item)):
+ return item
+ if item not in self.dataset:
+ return self.add(item)
+diff --git a/sos/cleaner/mappings/username_map.py b/sos/cleaner/mappings/username_map.py
+index 7ecccd7bc..ed6dc0912 100644
+--- a/sos/cleaner/mappings/username_map.py
++++ b/sos/cleaner/mappings/username_map.py
+@@ -24,7 +24,7 @@ class SoSUsernameMap(SoSMap):
+
+ def load_names_from_options(self, opt_names):
+ for name in opt_names:
+- if name not in self.dataset.keys():
++ if name and name not in self.dataset.keys():
+ self.add(name)
+
+ def sanitize_item(self, username):
+diff --git a/sos/cleaner/parsers/username_parser.py b/sos/cleaner/parsers/username_parser.py
+index 49640f7fd..2853c860f 100644
+--- a/sos/cleaner/parsers/username_parser.py
++++ b/sos/cleaner/parsers/username_parser.py
+@@ -55,7 +55,7 @@ def load_usernames_into_map(self, content):
+ user = line.split()[0]
+ except Exception:
+ continue
+- if user.lower() in self.skip_list:
++ if not user or user.lower() in self.skip_list:
+ continue
+ users.add(user)
+ for each in users:
+From ed618678fd3d07e68e1a430eb7d225a9701332e0 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Thu, 13 Jan 2022 13:52:34 -0500
+Subject: [PATCH] [clean,parsers] Build regex lists for static items only once
+
+For parsers such as the username and keyword parsers, we don't discover
+new items through parsing archives - these parsers use static lists
+determined before we begin the actual obfuscation process.
+
+As such, we can build a list of regexes for these static items once, and
+then reference those regexes during execution, rather than rebuilding
+the regex for each of these items for every obfuscation.
+
+For use cases where hundreds of items, e.g. hundreds of usernames, are
+being obfuscated this results in a significant performance increase.
+Individual per-file gains are minor - fractions of a second - however
+these gains build up over the course of the hundreds to thousands of
+files a typical archive can be expected to contain.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/cleaner/__init__.py | 9 +++++++++
+ sos/cleaner/parsers/__init__.py | 10 ++++++++++
+ sos/cleaner/parsers/keyword_parser.py | 15 ++++++++++-----
+ sos/cleaner/parsers/username_parser.py | 14 ++++++++------
+ tests/unittests/cleaner_tests.py | 1 +
+ 5 files changed, 38 insertions(+), 11 deletions(-)
+
+diff --git a/sos/cleaner/__init__.py b/sos/cleaner/__init__.py
+index 5686e2131..b76bef644 100644
+--- a/sos/cleaner/__init__.py
++++ b/sos/cleaner/__init__.py
+@@ -294,6 +294,7 @@ def execute(self):
+ # we have at least one valid target to obfuscate
+ self.completed_reports = []
+ self.preload_all_archives_into_maps()
++ self.generate_parser_item_regexes()
+ self.obfuscate_report_paths()
+
+ if not self.completed_reports:
+@@ -498,6 +499,14 @@ def _replace_obfuscated_archives(self):
+ shutil.move(archive.final_archive_path, dest)
+ archive.final_archive_path = dest_name
+
++ def generate_parser_item_regexes(self):
++ """For the parsers that use prebuilt lists of items, generate those
++ regexes now since all the parsers should be preloaded by the archive(s)
++ as well as being handed cmdline options and mapping file configuration.
++ """
++ for parser in self.parsers:
++ parser.generate_item_regexes()
++
+ def preload_all_archives_into_maps(self):
+ """Before doing the actual obfuscation, if we have multiple archives
+ to obfuscate then we need to preload each of them into the mappings
+diff --git a/sos/cleaner/parsers/__init__.py b/sos/cleaner/parsers/__init__.py
+index e62fd9384..6def863a6 100644
+--- a/sos/cleaner/parsers/__init__.py
++++ b/sos/cleaner/parsers/__init__.py
+@@ -46,9 +46,19 @@ class SoSCleanerParser():
+ map_file_key = 'unset'
+
+ def __init__(self, config={}):
++ self.regexes = {}
+ if self.map_file_key in config:
+ self.mapping.conf_update(config[self.map_file_key])
+
++ def generate_item_regexes(self):
++ """Generate regexes for items the parser will be searching for
++ repeatedly without needing to generate them for every file and/or line
++ we process
++
++ Not used by all parsers.
++ """
++ pass
++
+ def parse_line(self, line):
+ """This will be called for every line in every file we process, so that
+ every parser has a chance to scrub everything.
+diff --git a/sos/cleaner/parsers/keyword_parser.py b/sos/cleaner/parsers/keyword_parser.py
+index 694c6073a..362a1929e 100644
+--- a/sos/cleaner/parsers/keyword_parser.py
++++ b/sos/cleaner/parsers/keyword_parser.py
+@@ -9,6 +9,7 @@
+ # See the LICENSE file in the source distribution for further information.
+
+ import os
++import re
+
+ from sos.cleaner.parsers import SoSCleanerParser
+ from sos.cleaner.mappings.keyword_map import SoSKeywordMap
+@@ -33,16 +34,20 @@ def __init__(self, config, keywords=None, keyword_file=None):
+ # pre-generate an obfuscation mapping for each keyword
+ # this is necessary for cases where filenames are being
+ # obfuscated before or instead of file content
+- self.mapping.get(keyword)
++ self.mapping.get(keyword.lower())
+ self.user_keywords.append(keyword)
+ if keyword_file and os.path.exists(keyword_file):
+ with open(keyword_file, 'r') as kwf:
+ self.user_keywords.extend(kwf.read().splitlines())
+
++ def generate_item_regexes(self):
++ for kw in self.user_keywords:
++ self.regexes[kw] = re.compile(kw, re.I)
++
+ def parse_line(self, line):
+ count = 0
+- for keyword in sorted(self.user_keywords, reverse=True):
+- if keyword in line:
+- line = line.replace(keyword, self.mapping.get(keyword))
+- count += 1
++ for kwrd, reg in sorted(self.regexes.items(), key=len, reverse=True):
++ if reg.search(line):
++ line, _count = reg.subn(self.mapping.get(kwrd.lower()), line)
++ count += _count
+ return line, count
+diff --git a/sos/cleaner/parsers/username_parser.py b/sos/cleaner/parsers/username_parser.py
+index 3208a6557..49640f7fd 100644
+--- a/sos/cleaner/parsers/username_parser.py
++++ b/sos/cleaner/parsers/username_parser.py
+@@ -61,12 +61,14 @@ def load_usernames_into_map(self, content):
+ for each in users:
+ self.mapping.get(each)
+
++ def generate_item_regexes(self):
++ for user in self.mapping.dataset:
++ self.regexes[user] = re.compile(user, re.I)
++
+ def parse_line(self, line):
+ count = 0
+- for username in sorted(self.mapping.dataset.keys(), reverse=True):
+- _reg = re.compile(username, re.I)
+- if _reg.search(line):
+- line, count = _reg.subn(
+- self.mapping.get(username.lower()), line
+- )
++ for user, reg in sorted(self.regexes.items(), key=len, reverse=True):
++ if reg.search(line):
++ line, _count = reg.subn(self.mapping.get(user.lower()), line)
++ count += _count
+ return line, count
+diff --git a/tests/unittests/cleaner_tests.py b/tests/unittests/cleaner_tests.py
+index cb20772fd..b59eade9a 100644
+--- a/tests/unittests/cleaner_tests.py
++++ b/tests/unittests/cleaner_tests.py
+@@ -105,6 +105,7 @@ def setUp(self):
+ self.host_parser = SoSHostnameParser(config={}, opt_domains='foobar.com')
+ self.kw_parser = SoSKeywordParser(config={}, keywords=['foobar'])
+ self.kw_parser_none = SoSKeywordParser(config={})
++ self.kw_parser.generate_item_regexes()
+
+ def test_ip_parser_valid_ipv4_line(self):
+ line = 'foobar foo 10.0.0.1/24 barfoo bar'
+From 2ae16e0245e1b01b8547e507abb69c11871a8467 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 21 Feb 2022 14:37:09 -0500
+Subject: [PATCH] [sosnode] Handle downstream versioning for runtime option
+ check
+
+First, adds parsing and formatting for an sos installation's release
+version according to the loaded package manager for that node.
+
+Adds a fallback version check for 4.2-13 for RHEL downstreams that
+backport the `container-runtime` option into sos-4.2.
+
+Carry this in upstream to account for use cases where a workstation used
+to run `collect` from may be from a different stream than those used by
+cluster nodes.
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/collector/sosnode.py | 60 ++++++++++++++++++++++++++++++++++------
+ 1 file changed, 51 insertions(+), 9 deletions(-)
+
+diff --git a/sos/collector/sosnode.py b/sos/collector/sosnode.py
+index 7bbe0cd1..d9b998b0 100644
+--- a/sos/collector/sosnode.py
++++ b/sos/collector/sosnode.py
+@@ -275,21 +275,34 @@ class SosNode():
+ def _load_sos_info(self):
+ """Queries the node for information about the installed version of sos
+ """
++ ver = None
++ rel = None
+ if self.host.container_version_command is None:
+ pkg = self.host.package_manager.pkg_version(self.host.sos_pkg_name)
+ if pkg is not None:
+ ver = '.'.join(pkg['version'])
+- self.sos_info['version'] = ver
++ if pkg['release']:
++ rel = pkg['release']
++
+ else:
+ # use the containerized policy's command
+ pkgs = self.run_command(self.host.container_version_command,
+ use_container=True, need_root=True)
+ if pkgs['status'] == 0:
+- ver = pkgs['output'].strip().split('-')[1]
+- if ver:
+- self.sos_info['version'] = ver
+- else:
+- self.sos_info['version'] = None
++ _, ver, rel = pkgs['output'].strip().split('-')
++
++ if ver:
++ if len(ver.split('.')) == 2:
++ # safeguard against maintenance releases throwing off the
++ # comparison by LooseVersion
++ ver += '.0'
++ try:
++ ver += '-%s' % rel.split('.')[0]
++ except Exception as err:
++ self.log_debug("Unable to fully parse sos release: %s" % err)
++
++ self.sos_info['version'] = ver
++
+ if self.sos_info['version']:
+ self.log_info('sos version is %s' % self.sos_info['version'])
+ else:
+@@ -381,9 +394,37 @@ class SosNode():
+ """Checks to see if the sos installation on the node is AT LEAST the
+ given ver. This means that if the installed version is greater than
+ ver, this will still return True
++
++ :param ver: Version number we are trying to verify is installed
++ :type ver: ``str``
++
++ :returns: True if installed version is at least ``ver``, else False
++ :rtype: ``bool``
+ """
+- return self.sos_info['version'] is not None and \
+- LooseVersion(self.sos_info['version']) >= ver
++ def _format_version(ver):
++ # format the version we're checking to a standard form of X.Y.Z-R
++ try:
++ _fver = ver.split('-')[0]
++ _rel = ''
++ if '-' in ver:
++ _rel = '-' + ver.split('-')[-1].split('.')[0]
++ if len(_fver.split('.')) == 2:
++ _fver += '.0'
++
++ return _fver + _rel
++ except Exception as err:
++ self.log_debug("Unable to format '%s': %s" % (ver, err))
++ return ver
++
++ _ver = _format_version(ver)
++
++ try:
++ _node_ver = LooseVersion(self.sos_info['version'])
++ _test_ver = LooseVersion(_ver)
++ return _node_ver >= _test_ver
++ except Exception as err:
++ self.log_error("Error checking sos version: %s" % err)
++ return False
+
+ def is_installed(self, pkg):
+ """Checks if a given package is installed on the node"""
+@@ -587,7 +628,8 @@ class SosNode():
+ sos_opts.append('--cmd-timeout=%s'
+ % quote(str(self.opts.cmd_timeout)))
+
+- if self.check_sos_version('4.3'):
++ # handle downstream versions that backported this option
++ if self.check_sos_version('4.3') or self.check_sos_version('4.2-13'):
+ if self.opts.container_runtime != 'auto':
+ sos_opts.append(
+ "--container-runtime=%s" % self.opts.container_runtime
+--
+2.34.1
+
diff --git a/SOURCES/sos-bz2041855-virsh-in-foreground.patch b/SOURCES/sos-bz2041855-virsh-in-foreground.patch
new file mode 100644
index 0000000..66bca13
--- /dev/null
+++ b/SOURCES/sos-bz2041855-virsh-in-foreground.patch
@@ -0,0 +1,146 @@
+From 137abd394f64a63b6633949b5c81159af12038b7 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 14 Jan 2022 20:07:17 +0100
+Subject: [PATCH] [report] pass foreground argument to collect_cmd_output
+
+Related to: #2825
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/__init__.py | 12 +++++++++---
+ 1 file changed, 9 insertions(+), 3 deletions(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 98f163ab9..1bbdf28a4 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -1920,6 +1920,8 @@ class Plugin(object):
+ :param subdir: Subdir in plugin directory to save to
+ :param changes: Does this cmd potentially make a change
+ on the system?
++ :param foreground: Run the `cmd` in the foreground with a
++ TTY
+ :param tags: Add tags in the archive manifest
+ :param cmd_as_tag: Format command string to tag
+
+@@ -2145,7 +2147,8 @@ def collect_cmd_output(self, cmd, suggest_filename=None,
+ root_symlink=False, timeout=None,
+ stderr=True, chroot=True, runat=None, env=None,
+ binary=False, sizelimit=None, pred=None,
+- changes=False, subdir=None, tags=[]):
++ changes=False, foreground=False, subdir=None,
++ tags=[]):
+ """Execute a command and save the output to a file for inclusion in the
+ report, then return the results for further use by the plugin
+
+@@ -2188,6 +2191,9 @@ def collect_cmd_output(self, cmd, suggest_filename=None,
+ on the system?
+ :type changes: ``bool``
+
++ :param foreground: Run the `cmd` in the foreground with a TTY
++ :type foreground: ``bool``
++
+ :param tags: Add tags in the archive manifest
+ :type tags: ``str`` or a ``list`` of strings
+
+@@ -2206,8 +2212,8 @@ def collect_cmd_output(self, cmd, suggest_filename=None,
+ return self._collect_cmd_output(
+ cmd, suggest_filename=suggest_filename, root_symlink=root_symlink,
+ timeout=timeout, stderr=stderr, chroot=chroot, runat=runat,
+- env=env, binary=binary, sizelimit=sizelimit, subdir=subdir,
+- tags=tags
++ env=env, binary=binary, sizelimit=sizelimit, foreground=foreground,
++ subdir=subdir, tags=tags
+ )
+
+ def exec_cmd(self, cmd, timeout=None, stderr=True, chroot=True,
+From 747fef695e4ff08f320c5f03090bdefa7154c761 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 14 Jan 2022 20:10:22 +0100
+Subject: [PATCH] [virsh] Call virsh commands in the foreground / with a TTY
+
+In some virsh errors (like unable to connect to a hypervisor),
+the tool requires to communicate to TTY otherwise it can get stuck
+(when called via Popen with a timeout).
+
+Calling it on foreground prevents the stuck / waiting on cmd timeout.
+
+Resolves: #2825
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/virsh.py | 14 +++++++++-----
+ 1 file changed, 9 insertions(+), 5 deletions(-)
+
+diff --git a/sos/report/plugins/virsh.py b/sos/report/plugins/virsh.py
+index d6b7c16761..08f9a8488c 100644
+--- a/sos/report/plugins/virsh.py
++++ b/sos/report/plugins/virsh.py
+@@ -39,26 +39,30 @@ def setup(self):
+ ]
+
+ for subcmd in subcmds:
+- self.add_cmd_output('%s %s' % (cmd, subcmd))
++ self.add_cmd_output('%s %s' % (cmd, subcmd), foreground=True)
+
+ # get network, pool and nwfilter elements
+ for k in ['net', 'nwfilter', 'pool']:
+- k_list = self.collect_cmd_output('%s %s-list' % (cmd, k))
++ k_list = self.collect_cmd_output('%s %s-list' % (cmd, k),
++ foreground=True)
+ if k_list['status'] == 0:
+ k_lines = k_list['output'].splitlines()
+ # the 'Name' column position changes between virsh cmds
+ pos = k_lines[0].split().index('Name')
+ for j in filter(lambda x: x, k_lines[2:]):
+ n = j.split()[pos]
+- self.add_cmd_output('%s %s-dumpxml %s' % (cmd, k, n))
++ self.add_cmd_output('%s %s-dumpxml %s' % (cmd, k, n),
++ foreground=True)
+
+ # cycle through the VMs/domains list, ignore 2 header lines and latest
+ # empty line, and dumpxml domain name in 2nd column
+- domains_output = self.exec_cmd('%s list --all' % cmd)
++ domains_output = self.exec_cmd('%s list --all' % cmd, foreground=True)
+ if domains_output['status'] == 0:
+ domains_lines = domains_output['output'].splitlines()[2:]
+ for domain in filter(lambda x: x, domains_lines):
+ d = domain.split()[1]
+ for x in ['dumpxml', 'dominfo', 'domblklist']:
+- self.add_cmd_output('%s %s %s' % (cmd, x, d))
++ self.add_cmd_output('%s %s %s' % (cmd, x, d),
++ foreground=True)
++
+ # vim: et ts=4 sw=4
+From 9bc032129ec66766f07349dd115335f104888efa Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Wed, 26 Jan 2022 09:44:01 +0100
+Subject: [PATCH] [virsh] Catch parsing exception
+
+In case virsh output is malformed or missing 'Name' otherwise,
+catch parsing exception and continue in next for loop iteration.
+
+Resolves: #2836
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/virsh.py | 6 +++++-
+ 1 file changed, 5 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/virsh.py b/sos/report/plugins/virsh.py
+index 08f9a8488..2ce1df15c 100644
+--- a/sos/report/plugins/virsh.py
++++ b/sos/report/plugins/virsh.py
+@@ -48,7 +48,11 @@ def setup(self):
+ if k_list['status'] == 0:
+ k_lines = k_list['output'].splitlines()
+ # the 'Name' column position changes between virsh cmds
+- pos = k_lines[0].split().index('Name')
++ # catch the rare exceptions when 'Name' is not found
++ try:
++ pos = k_lines[0].split().index('Name')
++ except Exception:
++ continue
+ for j in filter(lambda x: x, k_lines[2:]):
+ n = j.split()[pos]
+ self.add_cmd_output('%s %s-dumpxml %s' % (cmd, k, n),
diff --git a/SOURCES/sos-bz2043104-foreman-tasks-msgpack.patch b/SOURCES/sos-bz2043104-foreman-tasks-msgpack.patch
new file mode 100644
index 0000000..900389c
--- /dev/null
+++ b/SOURCES/sos-bz2043104-foreman-tasks-msgpack.patch
@@ -0,0 +1,59 @@
+From 5634f7dd77eff821f37daa953fa86cc783d3b937 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Fri, 21 Jan 2022 16:27:33 +0100
+Subject: [PATCH] [foreman] Use psql-msgpack-decode wrapper for dynflow >= 1.6
+
+In dynflow >=1.6.3, dynflow* tables in postgres are encoded by
+msgpack which makes plain CSV dumps unreadable. In such a case,
+psql-msgpack-decode wrapper tool from dynflow-utils (of any
+version) must be used instead of the plain psql command.
+
+Resolves: #2830
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/foreman.py | 16 ++++++++++++----
+ 1 file changed, 12 insertions(+), 4 deletions(-)
+
+diff --git a/sos/report/plugins/foreman.py b/sos/report/plugins/foreman.py
+index 314a651d1..3fd80e6a8 100644
+--- a/sos/report/plugins/foreman.py
++++ b/sos/report/plugins/foreman.py
+@@ -244,8 +244,16 @@ def setup(self):
+ self.add_cmd_output(_cmd, suggest_filename=table, timeout=600,
+ sizelimit=100, env=self.env)
+
++ # dynflow* tables on dynflow >=1.6.3 are encoded and hence in that
++ # case, psql-msgpack-decode wrapper tool from dynflow-utils (any
++ # version) must be used instead of plain psql command
++ dynutils = self.is_installed('dynflow-utils')
+ for dyn in foremancsv:
+- _cmd = self.build_query_cmd(foremancsv[dyn], csv=True)
++ binary = "psql"
++ if dyn != 'foreman_tasks_tasks' and dynutils:
++ binary = "/usr/libexec/psql-msgpack-decode"
++ _cmd = self.build_query_cmd(foremancsv[dyn], csv=True,
++ binary=binary)
+ self.add_cmd_output(_cmd, suggest_filename=dyn, timeout=600,
+ sizelimit=100, env=self.env)
+
+@@ -270,7 +278,7 @@ def setup(self):
+ # collect http[|s]_proxy env.variables
+ self.add_env_var(["http_proxy", "https_proxy"])
+
+- def build_query_cmd(self, query, csv=False):
++ def build_query_cmd(self, query, csv=False, binary="psql"):
+ """
+ Builds the command needed to invoke the pgsql query as the postgres
+ user.
+@@ -281,8 +289,8 @@ def build_query_cmd(self, query, csv=False):
+ if csv:
+ query = "COPY (%s) TO STDOUT " \
+ "WITH (FORMAT 'csv', DELIMITER ',', HEADER)" % query
+- _dbcmd = "psql --no-password -h %s -p 5432 -U foreman -d foreman -c %s"
+- return _dbcmd % (self.dbhost, quote(query))
++ _dbcmd = "%s --no-password -h %s -p 5432 -U foreman -d foreman -c %s"
++ return _dbcmd % (binary, self.dbhost, quote(query))
+
+ def postproc(self):
+ self.do_path_regex_sub(
diff --git a/SOURCES/sos-bz2043488-ovn-proper-package-enablement.patch b/SOURCES/sos-bz2043488-ovn-proper-package-enablement.patch
new file mode 100644
index 0000000..16c48c4
--- /dev/null
+++ b/SOURCES/sos-bz2043488-ovn-proper-package-enablement.patch
@@ -0,0 +1,252 @@
+From 210b83e1d1164d29b1f6198675b8b596c4af8336 Mon Sep 17 00:00:00 2001
+From: Daniel Alvarez Sanchez <dalvarez@redhat.com>
+Date: Thu, 20 Jan 2022 12:58:44 +0100
+Subject: [PATCH] [ovn_central] Account for Red Hat ovn package naming
+
+Previous ovn packages were 'ovn2xxx' and now they have
+been renamed to 'ovn-2xxx'. This causes sos tool to not
+recognize that the packages are installed and it won't
+collect the relevant data.
+
+This patch is changing the match to be compatible
+with the previous and newer naming conventions.
+
+Signed-off-by: Daniel Alvarez Sanchez <dalvarez@redhat.com>
+---
+ sos/report/plugins/ovn_central.py | 2 +-
+ sos/report/plugins/ovn_host.py | 2 +-
+ 2 files changed, 2 insertions(+), 2 deletions(-)
+
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index ddbf288da..0f947d4c5 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -147,7 +147,7 @@ def setup(self):
+
+ class RedHatOVNCentral(OVNCentral, RedHatPlugin):
+
+- packages = ('openvswitch-ovn-central', 'ovn2.*-central', )
++ packages = ('openvswitch-ovn-central', 'ovn.*-central', )
+ ovn_sbdb_sock_path = '/var/run/openvswitch/ovnsb_db.ctl'
+
+
+diff --git a/sos/report/plugins/ovn_host.py b/sos/report/plugins/ovn_host.py
+index 78604a15a..25c38cccc 100644
+--- a/sos/report/plugins/ovn_host.py
++++ b/sos/report/plugins/ovn_host.py
+@@ -55,7 +55,7 @@ def check_enabled(self):
+
+ class RedHatOVNHost(OVNHost, RedHatPlugin):
+
+- packages = ('openvswitch-ovn-host', 'ovn2.*-host', )
++ packages = ('openvswitch-ovn-host', 'ovn.*-host', )
+
+
+ class DebianOVNHost(OVNHost, DebianPlugin, UbuntuPlugin):
+From 21fc376d97a5f74743e2b7cf7069349e874b979e Mon Sep 17 00:00:00 2001
+From: Hemanth Nakkina <hemanth.nakkina@canonical.com>
+Date: Fri, 4 Feb 2022 07:57:59 +0530
+Subject: [PATCH] [ovn-central] collect NB/SB ovsdb-server cluster status
+
+Add commands to collect cluster status of Northbound and
+Southbound ovsdb servers.
+
+Resolves: #2840
+
+Signed-off-by: Hemanth Nakkina hemanth.nakkina@canonical.com
+---
+ sos/report/plugins/ovn_central.py | 13 ++++++++++++-
+ 1 file changed, 12 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index 0f947d4c5..2f0438df3 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -84,6 +84,14 @@ def setup(self):
+ else:
+ self.add_copy_spec("/var/log/ovn/*.log")
+
++ # ovsdb nb/sb cluster status commands
++ ovsdb_cmds = [
++ 'ovs-appctl -t {} cluster/status OVN_Northbound'.format(
++ self.ovn_nbdb_sock_path),
++ 'ovs-appctl -t {} cluster/status OVN_Southbound'.format(
++ self.ovn_sbdb_sock_path),
++ ]
++
+ # Some user-friendly versions of DB output
+ nbctl_cmds = [
+ 'ovn-nbctl show',
+@@ -109,7 +117,8 @@ def setup(self):
+
+ self.add_database_output(nb_tables, nbctl_cmds, 'ovn-nbctl')
+
+- cmds = nbctl_cmds
++ cmds = ovsdb_cmds
++ cmds += nbctl_cmds
+
+ # Can only run sbdb commands if we are the leader
+ co = {'cmd': "ovs-appctl -t {} cluster/status OVN_Southbound".
+@@ -148,10 +157,12 @@ def setup(self):
+ class RedHatOVNCentral(OVNCentral, RedHatPlugin):
+
+ packages = ('openvswitch-ovn-central', 'ovn.*-central', )
++ ovn_nbdb_sock_path = '/var/run/openvswitch/ovnnb_db.ctl'
+ ovn_sbdb_sock_path = '/var/run/openvswitch/ovnsb_db.ctl'
+
+
+ class DebianOVNCentral(OVNCentral, DebianPlugin, UbuntuPlugin):
+
+ packages = ('ovn-central', )
++ ovn_nbdb_sock_path = '/var/run/ovn/ovnnb_db.ctl'
+ ovn_sbdb_sock_path = '/var/run/ovn/ovnsb_db.ctl'
+From d0f9d507b0ec63c9e8f3e5d7b6507d9d0f97c038 Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Tue, 15 Feb 2022 16:24:47 -0500
+Subject: [PATCH] [runtimes] Allow container IDs to be used with
+ `container_exists()`
+
+As container runtimes can interchange container names and container IDs,
+sos should also allow the use of container IDs when checking for the
+presence of a given container.
+
+In particular, this change unblocks the use of `Plugin.exec_cmd()` when
+used in conjunction with `Plugin.get_container_by_name()` to pick a
+container based on a provided regex that the container name may match.
+
+Related: #2856
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/policies/runtimes/__init__.py | 17 +++++++++++++++++
+ sos/report/plugins/__init__.py | 6 +++---
+ 2 files changed, 20 insertions(+), 3 deletions(-)
+
+diff --git a/sos/policies/runtimes/__init__.py b/sos/policies/runtimes/__init__.py
+index 5ac673544..d28373496 100644
+--- a/sos/policies/runtimes/__init__.py
++++ b/sos/policies/runtimes/__init__.py
+@@ -147,6 +147,23 @@ def get_volumes(self):
+ vols.append(ent[-1])
+ return vols
+
++ def container_exists(self, container):
++ """Check if a given container ID or name exists on the system from the
++ perspective of the container runtime.
++
++ Note that this will only check _running_ containers
++
++ :param container: The name or ID of the container
++ :type container: ``str``
++
++ :returns: True if the container exists, else False
++ :rtype: ``bool``
++ """
++ for _contup in self.containers:
++ if container in _contup:
++ return True
++ return False
++
+ def fmt_container_cmd(self, container, cmd, quotecmd):
+ """Format a command to run inside a container using the runtime
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index 2988be089..cc5cb65bc 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -2593,7 +2593,7 @@ def container_exists(self, name):
+ """If a container runtime is present, check to see if a container with
+ a given name is currently running
+
+- :param name: The name of the container to check presence of
++ :param name: The name or ID of the container to check presence of
+ :type name: ``str``
+
+ :returns: ``True`` if `name` exists, else ``False``
+@@ -2601,8 +2601,8 @@ def container_exists(self, name):
+ """
+ _runtime = self._get_container_runtime()
+ if _runtime is not None:
+- con = _runtime.get_container_by_name(name)
+- return con is not None
++ return (_runtime.container_exists(name) or
++ _runtime.get_container_by_name(name) is not None)
+ return False
+
+ def get_all_containers_by_regex(self, regex, get_all=False):
+
+From de9b020a72d1ceda39587db4c6d5acf72cd90da2 Mon Sep 17 00:00:00 2001
+From: Fernando Royo <froyo@redhat.com>
+Date: Tue, 15 Feb 2022 10:00:38 +0100
+Subject: [PATCH] [ovn_central] Rename container responsable of Red Hat
+ ovn_central plugin
+
+ovn_central plugin is running by container with
+name 'ovn-dbs-bundle*', a typo has been identified and
+this cause plugin ovn_central not enabled by default as it
+does not recognize any container responsible of this.
+
+This patch fix this container name match, searching schema db
+keeping backward compatibility with openvswitch.
+---
+ sos/report/plugins/ovn_central.py | 23 ++++++++++++-----------
+ 1 file changed, 12 insertions(+), 11 deletions(-)
+
+diff --git a/sos/report/plugins/ovn_central.py b/sos/report/plugins/ovn_central.py
+index 2f0438df..2f34bff0 100644
+--- a/sos/report/plugins/ovn_central.py
++++ b/sos/report/plugins/ovn_central.py
+@@ -24,7 +24,7 @@ class OVNCentral(Plugin):
+ short_desc = 'OVN Northd'
+ plugin_name = "ovn_central"
+ profiles = ('network', 'virt')
+- containers = ('ovs-db-bundle.*',)
++ containers = ('ovn-dbs-bundle.*',)
+
+ def get_tables_from_schema(self, filename, skip=[]):
+ if self._container_name:
+@@ -66,7 +66,7 @@ class OVNCentral(Plugin):
+ cmds.append('%s list %s' % (ovn_cmd, table))
+
+ def setup(self):
+- self._container_name = self.get_container_by_name('ovs-dbs-bundle.*')
++ self._container_name = self.get_container_by_name(self.containers[0])
+
+ ovs_rundir = os.environ.get('OVS_RUNDIR')
+ for pidfile in ['ovnnb_db.pid', 'ovnsb_db.pid', 'ovn-northd.pid']:
+@@ -110,12 +110,11 @@ class OVNCentral(Plugin):
+ 'ovn-sbctl get-connection',
+ ]
+
+- schema_dir = '/usr/share/openvswitch'
+-
+- nb_tables = self.get_tables_from_schema(self.path_join(
+- schema_dir, 'ovn-nb.ovsschema'))
+-
+- self.add_database_output(nb_tables, nbctl_cmds, 'ovn-nbctl')
++ # backward compatibility
++ for path in ['/usr/share/openvswitch', '/usr/share/ovn']:
++ nb_tables = self.get_tables_from_schema(self.path_join(
++ path, 'ovn-nb.ovsschema'))
++ self.add_database_output(nb_tables, nbctl_cmds, 'ovn-nbctl')
+
+ cmds = ovsdb_cmds
+ cmds += nbctl_cmds
+@@ -125,9 +124,11 @@ class OVNCentral(Plugin):
+ format(self.ovn_sbdb_sock_path),
+ "output": "Leader: self"}
+ if self.test_predicate(self, pred=SoSPredicate(self, cmd_outputs=co)):
+- sb_tables = self.get_tables_from_schema(self.path_join(
+- schema_dir, 'ovn-sb.ovsschema'), ['Logical_Flow'])
+- self.add_database_output(sb_tables, sbctl_cmds, 'ovn-sbctl')
++ # backward compatibility
++ for path in ['/usr/share/openvswitch', '/usr/share/ovn']:
++ sb_tables = self.get_tables_from_schema(self.path_join(
++ path, 'ovn-sb.ovsschema'), ['Logical_Flow'])
++ self.add_database_output(sb_tables, sbctl_cmds, 'ovn-sbctl')
+ cmds += sbctl_cmds
+
+ # If OVN is containerized, we need to run the above commands inside
+--
+2.34.1
+
diff --git a/SOURCES/sos-bz2054883-plugopt-logging-effective-opts.patch b/SOURCES/sos-bz2054883-plugopt-logging-effective-opts.patch
new file mode 100644
index 0000000..f8e7ed3
--- /dev/null
+++ b/SOURCES/sos-bz2054883-plugopt-logging-effective-opts.patch
@@ -0,0 +1,94 @@
+From 5824cd5d3bddf39e0382d568419e2453abc93d8a Mon Sep 17 00:00:00 2001
+From: Jake Hunsaker <jhunsake@redhat.com>
+Date: Mon, 30 Aug 2021 15:09:07 -0400
+Subject: [PATCH] [options] Fix logging on plugopts in effective sos command
+
+First, provide a special-case handling for plugin options specified in
+sos.conf in `SoSOptions.to_args().has_value()` that allows for plugin
+options to be included in the "effective options now" log message.
+
+Second, move the logging of said message (and thus the merging of
+preset options, if used), to being _prior_ to the loading of plugin
+options.
+
+Combined, plugin options specified in sos.conf will now be logged
+properly and this logging will occur before we set (and log the setting
+of) those options.
+
+Resolves: #2663
+
+Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
+---
+ sos/options.py | 2 ++
+ sos/report/__init__.py | 30 ++++++++++++++++--------------
+ 2 files changed, 18 insertions(+), 14 deletions(-)
+
+diff --git a/sos/options.py b/sos/options.py
+index a014a022..7bea3ffc 100644
+--- a/sos/options.py
++++ b/sos/options.py
+@@ -281,6 +281,8 @@ class SoSOptions():
+ null_values = ("False", "None", "[]", '""', "''", "0")
+ if not value or value in null_values:
+ return False
++ if name == 'plugopts' and value:
++ return True
+ if name in self.arg_defaults:
+ if str(value) == str(self.arg_defaults[name]):
+ return False
+diff --git a/sos/report/__init__.py b/sos/report/__init__.py
+index b0159e5b..82484f1d 100644
+--- a/sos/report/__init__.py
++++ b/sos/report/__init__.py
+@@ -925,20 +925,6 @@ class SoSReport(SoSComponent):
+ self._exit(1)
+
+ def setup(self):
+- # Log command line options
+- msg = "[%s:%s] executing 'sos %s'"
+- self.soslog.info(msg % (__name__, "setup", " ".join(self.cmdline)))
+-
+- # Log active preset defaults
+- preset_args = self.preset.opts.to_args()
+- msg = ("[%s:%s] using '%s' preset defaults (%s)" %
+- (__name__, "setup", self.preset.name, " ".join(preset_args)))
+- self.soslog.info(msg)
+-
+- # Log effective options after applying preset defaults
+- self.soslog.info("[%s:%s] effective options now: %s" %
+- (__name__, "setup", " ".join(self.opts.to_args())))
+-
+ self.ui_log.info(_(" Setting up plugins ..."))
+ for plugname, plug in self.loaded_plugins:
+ try:
+@@ -1386,11 +1372,27 @@ class SoSReport(SoSComponent):
+ self.report_md.add_list('disabled_plugins', self.opts.skip_plugins)
+ self.report_md.add_section('plugins')
+
++ def _merge_preset_options(self):
++ # Log command line options
++ msg = "[%s:%s] executing 'sos %s'"
++ self.soslog.info(msg % (__name__, "setup", " ".join(self.cmdline)))
++
++ # Log active preset defaults
++ preset_args = self.preset.opts.to_args()
++ msg = ("[%s:%s] using '%s' preset defaults (%s)" %
++ (__name__, "setup", self.preset.name, " ".join(preset_args)))
++ self.soslog.info(msg)
++
++ # Log effective options after applying preset defaults
++ self.soslog.info("[%s:%s] effective options now: %s" %
++ (__name__, "setup", " ".join(self.opts.to_args())))
++
+ def execute(self):
+ try:
+ self.policy.set_commons(self.get_commons())
+ self.load_plugins()
+ self._set_all_options()
++ self._merge_preset_options()
+ self._set_tunables()
+ self._check_for_unknown_plugins()
+ self._set_plugin_options()
+--
+2.34.1
+
diff --git a/SOURCES/sos-bz2055548-honour-plugins-timeout-hardcoded.patch b/SOURCES/sos-bz2055548-honour-plugins-timeout-hardcoded.patch
new file mode 100644
index 0000000..3adde40
--- /dev/null
+++ b/SOURCES/sos-bz2055548-honour-plugins-timeout-hardcoded.patch
@@ -0,0 +1,39 @@
+From 7069e99d1c5c443f96a98a7ed6db67fa14683e67 Mon Sep 17 00:00:00 2001
+From: Pavel Moravec <pmoravec@redhat.com>
+Date: Thu, 17 Feb 2022 09:14:15 +0100
+Subject: [PATCH] [report] Honor plugins' hardcoded plugin_timeout
+
+Currently, plugin's plugin_timeout hardcoded default is superseded by
+whatever --plugin-timeout value, even when this option is not used and
+we eval it to TIMEOUT_DEFAULT.
+
+In this case of not setting --plugin-timeout either -k plugin.timeout,
+honour plugin's plugin_timeout instead.
+
+Resolves: #2863
+Closes: #2864
+
+Signed-off-by: Pavel Moravec <pmoravec@redhat.com>
+---
+ sos/report/plugins/__init__.py | 5 ++++-
+ 1 file changed, 4 insertions(+), 1 deletion(-)
+
+diff --git a/sos/report/plugins/__init__.py b/sos/report/plugins/__init__.py
+index cc5cb65b..336b4d22 100644
+--- a/sos/report/plugins/__init__.py
++++ b/sos/report/plugins/__init__.py
+@@ -636,7 +636,10 @@ class Plugin():
+ if opt_timeout is None:
+ _timeout = own_timeout
+ elif opt_timeout is not None and own_timeout == -1:
+- _timeout = int(opt_timeout)
++ if opt_timeout == TIMEOUT_DEFAULT:
++ _timeout = default_timeout
++ else:
++ _timeout = int(opt_timeout)
+ elif opt_timeout is not None and own_timeout > -1:
+ _timeout = own_timeout
+ else:
+--
+2.34.1
+
diff --git a/SPECS/sos.spec b/SPECS/sos.spec
new file mode 100644
index 0000000..e468c18
--- /dev/null
+++ b/SPECS/sos.spec
@@ -0,0 +1,314 @@
+%{!?python_sitelib: %define python_sitelib %(%{__python} -c "from distutils.sysconfig import get_python_lib; print get_python_lib()")}
+
+%global auditversion 0.3
+
+Summary: A set of tools to gather troubleshooting information from a system
+Name: sos
+Version: 4.2
+Release: 15%{?dist}
+Group: Applications/System
+Source0: https://github.com/sosreport/sos/archive/%{version}/sos-%{version}.tar.gz
+Source1: sos-audit-%{auditversion}.tgz
+License: GPLv2+
+BuildArch: noarch
+Url: https://github.com/sosreport/sos
+BuildRequires: python3-devel
+BuildRequires: gettext
+Requires: libxml2-python3
+#Requires: python3-rpm
+Requires: tar
+Requires: bzip2
+Requires: xz
+Recommends: python3-pexpect
+Conflicts: vdsm < 4.40
+Obsoletes: sos-collector <= 1.9
+Recommends: python3-pexpect
+Recommends: python3-requests
+Patch1: sos-bz1869561-cpuX-individual-sizelimits.patch
+Patch2: sos-bz2011533-unpackaged-recursive-symlink.patch
+Patch3: sos-bz2011534-opacapture-under-allow-system-changes.patch
+Patch4: sos-bz2011535-kernel-psi.patch
+Patch5: sos-bz2011538-iptables-save-under-nf_tables-kmod.patch
+Patch6: sos-bz2011537-estimate-only-option.patch
+Patch7: sos-bz2011536-iptables-based-on-ntf.patch
+Patch8: sos-bz2011507-foreman-puma-status.patch
+Patch9: sos-bz2012858-dryrun-uncaught-exception.patch
+Patch10: sos-bz2019697-openvswitch-offline-analysis.patch
+Patch11: sos-bz2012859-plugin-timeout-unhandled-exception.patch
+Patch12: sos-bz2023481-plugin-timeouts-proper-handling.patch
+Patch13: sos-bz2020778-filter-namespace-per-pattern.patch
+Patch14: sos-bz2024893-cleaner-hostnames-improvements.patch
+Patch15: sos-bz2025611-RHTS-api-change.patch
+Patch16: sos-bz2034001-nvidia-GPU-info.patch
+Patch17: sos-bz2031777-rhui-logs.patch
+Patch18: sos-bz2037350-ocp-backports.patch
+Patch19: sos-bz2043104-foreman-tasks-msgpack.patch
+Patch20: sos-bz2041855-virsh-in-foreground.patch
+Patch21: sos-bz2043488-ovn-proper-package-enablement.patch
+Patch22: sos-bz2054883-plugopt-logging-effective-opts.patch
+Patch23: sos-bz2055548-honour-plugins-timeout-hardcoded.patch
+
+%description
+Sos is a set of tools that gathers information about system
+hardware and configuration. The information can then be used for
+diagnostic purposes and debugging. Sos is commonly used to help
+support technicians and developers.
+
+%prep
+%setup -qn %{name}-%{version}
+%setup -T -D -a1 -q
+%patch1 -p1
+%patch2 -p1
+%patch3 -p1
+%patch4 -p1
+%patch5 -p1
+%patch6 -p1
+%patch7 -p1
+%patch8 -p1
+%patch9 -p1
+%patch10 -p1
+%patch11 -p1
+%patch12 -p1
+%patch13 -p1
+%patch14 -p1
+%patch15 -p1
+%patch16 -p1
+%patch17 -p1
+%patch18 -p1
+%patch19 -p1
+%patch20 -p1
+%patch21 -p1
+%patch22 -p1
+%patch23 -p1
+
+%build
+%py3_build
+
+%install
+%py3_install '--install-scripts=%{_sbindir}'
+
+install -d -m 755 %{buildroot}%{_sysconfdir}/%{name}
+install -d -m 700 %{buildroot}%{_sysconfdir}/%{name}/cleaner
+install -d -m 755 %{buildroot}%{_sysconfdir}/%{name}/presets.d
+install -d -m 755 %{buildroot}%{_sysconfdir}/%{name}/groups.d
+install -d -m 755 %{buildroot}%{_sysconfdir}/%{name}/extras.d
+install -m 644 %{name}.conf %{buildroot}%{_sysconfdir}/%{name}/%{name}.conf
+
+rm -rf %{buildroot}/usr/config/
+
+%find_lang %{name} || echo 0
+
+cd %{name}-audit-%{auditversion}
+DESTDIR=%{buildroot} ./install.sh
+cd ..
+
+%files -f %{name}.lang
+%{_sbindir}/sos
+%{_sbindir}/sosreport
+%{_sbindir}/sos-collector
+#%dir /etc/sos/cleaner
+%dir /etc/sos/presets.d
+%dir /etc/sos/extras.d
+%dir /etc/sos/groups.d
+%{python3_sitelib}/*
+%{_mandir}/man1/*
+%{_mandir}/man5/sos.conf.5.gz
+%doc AUTHORS README.md
+%license LICENSE
+%config(noreplace) %{_sysconfdir}/sos/sos.conf
+%config(noreplace) %{_sysconfdir}/sos/cleaner
+
+
+%package audit
+Summary: Audit use of some commands for support purposes
+License: GPLv2+
+Group: Application/System
+
+%description audit
+
+Sos-audit provides configuration files for the Linux Auditing System
+to track the use of some commands capable of changing the configuration
+of the system. Currently storage and filesystem commands are audited.
+
+%post audit
+%{_sbindir}/sos-audit.sh
+
+%files audit
+%defattr(755,root,root,-)
+%{_sbindir}/sos-audit.sh
+%defattr(644,root,root,-)
+%config(noreplace) %{_sysconfdir}/sos/sos-audit.conf
+%defattr(444,root,root,-)
+%{_prefix}/lib/sos/audit/*
+%{_mandir}/man5/sos-audit.conf.5.gz
+%{_mandir}/man8/sos-audit.sh.8.gz
+%ghost /etc/audit/rules.d/40-sos-filesystem.rules
+%ghost /etc/audit/rules.d/40-sos-storage.rules
+
+
+%changelog
+* Wed Feb 23 2022 Pavel Moravec <pmoravec@redhat.com> = 4.2-15
+- [sosnode] Handle downstream versioning for runtime option
+ Resolves: bz2037350
+- [options] Fix logging on plugopts in effective sos command
+ Resolves: bz2054883
+- [report] Honor plugins' hardcoded plugin_timeout
+ Resolves: bz2055548
+- [policies] Set fallback to None sysroot, don't chroot to '/'
+ Resolves: bz2011537
+- [ovn_central] Rename container responsable of Red Hat
+ Resolves: bz2043488
+
+* Wed Jan 26 2022 Pavel Moravec <pmoravec@redhat.com> = 4.2-13
+- [virsh] Catch parsing exception
+ Resolves: bz2041855
+
+* Tue Jan 25 2022 Pavel Moravec <pmoravec@redhat.com> = 4.2-12
+- [foreman] Use psql-msgpack-decode wrapper for dynflow >= 1.6
+ Resolves: bz2043104
+- [virsh] Call virsh commands in the foreground / with a TTY
+ Resolves: bz2041855
+- [ovn_central] Account for Red Hat ovn package naming
+ Resolves: bz2043488
+- [clean,parsers] Build regex lists for static items only once
+ Resolves: bz2037350
+
+* Mon Jan 10 2022 Pavel Moravec <pmoravec@redhat.com> = 4.2-11
+- [report] Add journal logs for NetworkManager plugin
+ Resolves: bz2037350
+
+* Fri Jan 07 2022 Pavel Moravec <pmoravec@redhat.com> = 4.2-9
+- add oc transport, backport various PRs for OCP
+ Resolves: bz2037350
+- [report] Provide better warning about estimate-mode
+ Resolves: bz2011537
+- [hostname] Fix loading and detection of long base domains
+ Resolves: bz2024893
+
+* Sun Dec 19 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-8
+- [rhui] New log folder
+ Resolves: bz2031777
+- nvidia]:Patch to update nvidia plugin for GPU info
+ Resolves: bz2034001
+- [hostname] Fix edge case for new hosts in a known subdomain
+ Resolves: bz2024893
+
+* Wed Dec 08 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-7
+- [hostname] Simplify case matching for domains
+ Resolves: bz2024893
+
+* Tue Nov 30 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-6
+- [redhat] Fix broken URI to upload to customer portal
+ Resolves: bz2025611
+
+* Mon Nov 22 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-5
+- [clean,hostname_parser] Source /etc/hosts for obfuscation
+ Resolves: bz2024893
+- [clean, hostname] Fix unintentionally case sensitive
+ Resolves: bz2024892
+- [redhat] update SFTP API version to v2
+ Resolves: bz2025611
+
+* Tue Nov 16 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-4
+- [report] Calculate sizes of dirs, symlinks and manifest in
+ Resolves: bz2011537
+- [report] shutdown threads for timeouted plugins
+ Resolves: bz2012859
+- [report] fix filter_namespace per pattern
+ Resolves: bz2020778
+- Ensure specific plugin timeouts are only set
+ Resolves: bz2023481
+
+* Wed Nov 03 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-2
+- [firewall_tables] call iptables -t <table> based on nft
+ Resolves: bz2011536
+- [report] Count with sos_logs and sos_reports in
+ Resolves: bz2011537
+- [foreman] Collect puma status and stats
+ Resolves: bz2011507
+- [report] Overwrite pred=None before refering predicate
+ Resolves: bz2012858
+- [openvswitch] add commands for offline analysis
+ Resolves: bz2019697
+
+* Wed Oct 06 2021 Pavel Moravec <pmoravec@redhat.com> = 4.2-1
+- Rebase on upstream 4.2
+ Resolves: bz1998134
+- [report] Implement --estimate-only
+ Resolves: bz2011537
+- [omnipath_client] Opacapture to run only with allow changes
+ Resolves: bz2011534
+- [unpackaged] deal with recursive loop of symlinks properly
+ Resolves: bz2011533
+- [networking] prevent iptables-save commands to load nf_tables
+ Resolves: bz2011538
+- [kernel] Capture Pressure Stall Information
+ Resolves: bz2011535
+- [processor] Apply sizelimit to /sys/devices/system/cpu/cpuX
+ Resolves: bz1869561
+
+* Wed Aug 11 2021 Pavel Moravec <pmoravec@redhat.com> = 4.1-8
+- [report,collect] unify --map-file arguments
+ Resolves: bz1985985
+- [rhui] add new plugin for RHUI 4
+ Resolves: bz1992859
+- [username parser] Load usernames from `last` for LDAP users
+ Resolves: bz1992861
+
+* Tue Aug 10 2021 Mohan Boddu <mboddu@redhat.com> - 4.1-7
+- Rebuilt for IMA sigs, glibc 2.34, aarch64 flags
+ Related: rhbz#1991688
+
+* Tue Jul 27 2021 Pavel Moravec <pmoravec@redhat.com> - 4.1-6
+- [networking] collect also tc filter show ingress
+ Resolves: bz1985976
+- [cleaner] Only skip packaging-based files for the IP parser
+ Resolves: bz1985982
+- [sssd] sssd plugin when sssd-common
+ Resolves: bz1967718
+- Various OCP/cluster/cleanup enhancements
+ Resolves: bz1985983
+- [options] allow variant option names in config file
+ Resolves: bz1985985
+- [plugins] Set default predicate instead of None
+ Resolves: bz1938874
+- [MigrationResults] collect info about conversions and
+ Resolves: bz1959779
+
+* Wed Jun 02 2021 Pavel Moravec <pmoravec@redhat.com> - 4.1-4
+- [archive] skip copying SELinux context for /proc and /sys everytime
+ Resolves: bz1965002
+- Load maps from all archives before obfuscation
+ Resolves: bz1967110
+- Multiple fixes in man pages
+ Resolves: bz1967111
+- [ds] Mask password and encryption keys in ldif files
+ Resolves: bz1967112
+- [report] add --cmd-timeout option
+ Resolves: bz1967113
+- [cups] Add gathering cups-browsed logs
+ Resolves: bz1967114
+- [sssd] Collect memory cache / individual logfiles
+ Resolves: bz1967115
+- Collect ibmvNIC dynamic_debugs
+ Resolves: bz1967116
+- [pulpcore] add plugin for pulp-3
+ Resolves: bz1967117
+- [saphana] remove redundant unused argument of get_inst_info
+ Resolves: bz1967118
+- [networking] Add nstat command support
+ Resolves: bz1967119
+- [snapper] add a new plugin
+ Resolves: bz1967120
+
+* Fri Apr 16 2021 Mohan Boddu <mboddu@redhat.com> - 4.1-4
+- Rebuilt for RHEL 9 BETA on Apr 15th 2021. Related: rhbz#1947937
+
+* Thu Apr 01 2021 Pavel Moravec <pmoravec@redhat.com> - 4.1-3
+- adding sos-audit
+- [gluster] Add glusterd public keys and status files
+ Resolves: bz1925419
+
+* Wed Mar 10 2021 Sandro Bonazzola <sbonazzo@redhat.com> - 4.1-1
+- Rebase to 4.1
+