diff --git a/.corosync.metadata b/.corosync.metadata index 7942f9e..c756eef 100644 --- a/.corosync.metadata +++ b/.corosync.metadata @@ -1,2 +1,2 @@ -dea662d9ae9a624196e3f116a29cece2eb979616 SOURCES/corosync-3.0.3.tar.gz +761fe353b2cbead7a8572bfb6b84fe5d2fc8d9d6 SOURCES/corosync-3.1.0.tar.gz 87cb5c01333298d106fc0022f5ac855205da0ae6 SOURCES/spausedd-20190807.tar.gz diff --git a/.gitignore b/.gitignore index 0833995..9c5438b 100644 --- a/.gitignore +++ b/.gitignore @@ -1,2 +1,2 @@ -SOURCES/corosync-3.0.3.tar.gz +SOURCES/corosync-3.1.0.tar.gz SOURCES/spausedd-20190807.tar.gz diff --git a/SOURCES/bz1780137-1-votequorum-Ignore-the-icmap_get_-return-value.patch b/SOURCES/bz1780137-1-votequorum-Ignore-the-icmap_get_-return-value.patch deleted file mode 100644 index d337c86..0000000 --- a/SOURCES/bz1780137-1-votequorum-Ignore-the-icmap_get_-return-value.patch +++ /dev/null @@ -1,73 +0,0 @@ -From cddd62f972bca276c934e58f08da84071cec1ddb Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Mon, 25 Nov 2019 18:21:52 +0100 -Subject: [PATCH] votequorum: Ignore the icmap_get_* return value - -Express intention to ignore icmap_get_* return -value and rely on default behavior of not changing the output -parameter on error. - -Signed-off-by: Jan Friesse ---- - exec/votequorum.c | 22 +++++++++++----------- - 1 file changed, 11 insertions(+), 11 deletions(-) - -diff --git a/exec/votequorum.c b/exec/votequorum.c -index f78b3f9..e1d7e73 100644 ---- a/exec/votequorum.c -+++ b/exec/votequorum.c -@@ -1271,10 +1271,10 @@ static char *votequorum_readconfig(int runtime) - /* - * gather basic data here - */ -- icmap_get_uint32("quorum.expected_votes", &expected_votes); -+ (void)icmap_get_uint32("quorum.expected_votes", &expected_votes); - have_nodelist = votequorum_read_nodelist_configuration(&node_votes, &node_count, &node_expected_votes); - have_qdevice = votequorum_qdevice_is_configured(&qdevice_votes); -- icmap_get_uint8("quorum.two_node", &two_node); -+ (void)icmap_get_uint8("quorum.two_node", &two_node); - - /* - * do config verification and enablement -@@ -1319,13 +1319,13 @@ static char *votequorum_readconfig(int runtime) - wait_for_all = 1; - } - -- icmap_get_uint8("quorum.allow_downscale", &allow_downscale); -- icmap_get_uint8("quorum.wait_for_all", &wait_for_all); -- icmap_get_uint8("quorum.last_man_standing", &last_man_standing); -- icmap_get_uint32("quorum.last_man_standing_window", &last_man_standing_window); -- icmap_get_uint8("quorum.expected_votes_tracking", &ev_tracking); -- icmap_get_uint8("quorum.auto_tie_breaker", &atb); -- icmap_get_string("quorum.auto_tie_breaker_node", &atb_string); -+ (void)icmap_get_uint8("quorum.allow_downscale", &allow_downscale); -+ (void)icmap_get_uint8("quorum.wait_for_all", &wait_for_all); -+ (void)icmap_get_uint8("quorum.last_man_standing", &last_man_standing); -+ (void)icmap_get_uint32("quorum.last_man_standing_window", &last_man_standing_window); -+ (void)icmap_get_uint8("quorum.expected_votes_tracking", &ev_tracking); -+ (void)icmap_get_uint8("quorum.auto_tie_breaker", &atb); -+ (void)icmap_get_string("quorum.auto_tie_breaker_node", &atb_string); - - /* auto_tie_breaker defaults to LOWEST */ - if (atb) { -@@ -1517,7 +1517,7 @@ static char *votequorum_readconfig(int runtime) - us->expected_votes = node_expected_votes; - } else { - us->votes = 1; -- icmap_get_uint32("quorum.votes", &us->votes); -+ (void)icmap_get_uint32("quorum.votes", &us->votes); - } - - if (expected_votes) { -@@ -1568,7 +1568,7 @@ static void votequorum_refresh_config( - return ; - } - -- icmap_get_uint8("quorum.cancel_wait_for_all", &cancel_wfa); -+ (void)icmap_get_uint8("quorum.cancel_wait_for_all", &cancel_wfa); - if (strcmp(key_name, "quorum.cancel_wait_for_all") == 0 && - cancel_wfa >= 1) { - icmap_set_uint8("quorum.cancel_wait_for_all", 0); --- -1.8.3.1 - diff --git a/SOURCES/bz1780137-2-votequorum-Reflect-runtime-change-of-2Node-to-WFA.patch b/SOURCES/bz1780137-2-votequorum-Reflect-runtime-change-of-2Node-to-WFA.patch deleted file mode 100644 index 0410773..0000000 --- a/SOURCES/bz1780137-2-votequorum-Reflect-runtime-change-of-2Node-to-WFA.patch +++ /dev/null @@ -1,80 +0,0 @@ -From 8ce65bf951bc1e5b2d64b60ea027fbdc551d4fc8 Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Thu, 16 Jan 2020 15:43:59 +0100 -Subject: [PATCH] votequorum: Reflect runtime change of 2Node to WFA - -When 2Node mode is set, WFA is also set unless WFA is configured -explicitly. This behavior was not reflected on runtime change, so -restarted corosync behavior was different (WFA not set). Also when -cluster is reduced from 3 nodes to 2 nodes during runtime, WFA was not -set, what may result in two quorate partitions. - -Solution is to set WFA depending on 2Node when WFA -is not explicitly configured. - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - exec/votequorum.c | 24 +++++++++++++++++++----- - 1 file changed, 19 insertions(+), 5 deletions(-) - -diff --git a/exec/votequorum.c b/exec/votequorum.c -index 0cde8f8..52424fa 100644 ---- a/exec/votequorum.c -+++ b/exec/votequorum.c -@@ -80,6 +80,7 @@ static uint8_t two_node = 0; - - static uint8_t wait_for_all = 0; - static uint8_t wait_for_all_status = 0; -+static uint8_t wait_for_all_autoset = 0; /* Wait for all is not set explicitly and follows two_node */ - - static enum {ATB_NONE, ATB_LOWEST, ATB_HIGHEST, ATB_LIST} auto_tie_breaker = ATB_NONE, initial_auto_tie_breaker = ATB_NONE; - static int lowest_node_id = -1; -@@ -1315,12 +1316,10 @@ static char *votequorum_readconfig(int runtime) - * Enable special features - */ - if (!runtime) { -- if (two_node) { -- wait_for_all = 1; -- } -- - (void)icmap_get_uint8("quorum.allow_downscale", &allow_downscale); -- (void)icmap_get_uint8("quorum.wait_for_all", &wait_for_all); -+ if (icmap_get_uint8("quorum.wait_for_all", &wait_for_all) != CS_OK) { -+ wait_for_all_autoset = 1; -+ } - (void)icmap_get_uint8("quorum.last_man_standing", &last_man_standing); - (void)icmap_get_uint32("quorum.last_man_standing_window", &last_man_standing_window); - (void)icmap_get_uint8("quorum.expected_votes_tracking", &ev_tracking); -@@ -1361,6 +1360,15 @@ static char *votequorum_readconfig(int runtime) - - } - -+ /* -+ * Changing of wait_for_all during runtime is not supported, but changing of two_node is -+ * and two_node may set wfa if not configured explicitly. It is safe to unset it -+ * (or set it back) when two_node changes. -+ */ -+ if (wait_for_all_autoset) { -+ wait_for_all = two_node; -+ } -+ - /* two_node and auto_tie_breaker are not compatible as two_node uses - * a fence race to decide quorum whereas ATB decides based on node id - */ -@@ -1540,6 +1548,12 @@ static char *votequorum_readconfig(int runtime) - update_two_node(); - if (wait_for_all) { - update_wait_for_all_status(1); -+ } else if (wait_for_all_autoset && wait_for_all_status) { -+ /* -+ * Reset wait for all status for consistency when wfa is auto-unset by 2node. -+ * wait_for_all_status would be ignored by are_we_quorate anyway. -+ */ -+ update_wait_for_all_status(0); - } - - out: --- -1.8.3.1 - diff --git a/SOURCES/bz1791792-1-man-move-cmap_keys-man-page-from-section-8-to-7.patch b/SOURCES/bz1791792-1-man-move-cmap_keys-man-page-from-section-8-to-7.patch deleted file mode 100644 index 4141cbf..0000000 --- a/SOURCES/bz1791792-1-man-move-cmap_keys-man-page-from-section-8-to-7.patch +++ /dev/null @@ -1,901 +0,0 @@ -From f1d36307e524f9440733f0b01a9fc627a0e1cac7 Mon Sep 17 00:00:00 2001 -From: =?UTF-8?q?Ferenc=20W=C3=A1gner?= -Date: Sat, 4 Jan 2020 13:38:08 +0100 -Subject: [PATCH] man: move cmap_keys man page from section 8 to 7 -MIME-Version: 1.0 -Content-Type: text/plain; charset=UTF-8 -Content-Transfer-Encoding: 8bit - -Section 8 is for "System administration commands", 7 is "Miscellaneous". - -Signed-off-by: Ferenc Wágner -Reviewed-by: Jan Friesse ---- - corosync.spec.in | 2 +- - man/Makefile.am | 2 +- - man/cmap_keys.7 | 397 +++++++++++++++++++++++++++++++++++++++++++++++++ - man/cmap_keys.8 | 397 ------------------------------------------------- - man/cmap_overview.3 | 4 +- - man/corosync-cmapctl.8 | 2 +- - man/index.html | 2 +- - 7 files changed, 403 insertions(+), 403 deletions(-) - create mode 100644 man/cmap_keys.7 - delete mode 100644 man/cmap_keys.8 - -diff --git a/corosync.spec.in b/corosync.spec.in -index c06675d..8ac3757 100644 ---- a/corosync.spec.in -+++ b/corosync.spec.in -@@ -217,7 +217,7 @@ fi - %{_mandir}/man8/corosync-quorumtool.8* - %{_mandir}/man5/corosync.conf.5* - %{_mandir}/man5/votequorum.5* --%{_mandir}/man8/cmap_keys.8* -+%{_mandir}/man7/cmap_keys.7* - - # library - # -diff --git a/man/Makefile.am b/man/Makefile.am -index 2ef5dcd..92a76ed 100644 ---- a/man/Makefile.am -+++ b/man/Makefile.am -@@ -140,7 +140,7 @@ dist_man_MANS = corosync.conf.5 \ - votequorum_overview.3 \ - sam_overview.3 \ - cmap_overview.3 \ -- cmap_keys.8 -+ cmap_keys.7 - - if BUILD_VQSIM - dist_man_MANS += $(corosync_vqsim_man) -diff --git a/man/cmap_keys.7 b/man/cmap_keys.7 -new file mode 100644 -index 0000000..6bc04fe ---- /dev/null -+++ b/man/cmap_keys.7 -@@ -0,0 +1,397 @@ -+.\"/* -+.\" * Copyright (c) 2012-2018 Red Hat, Inc. -+.\" * -+.\" * All rights reserved. -+.\" * -+.\" * Author: Jan Friesse (jfriesse@redhat.com) -+.\" * -+.\" * This software licensed under BSD license, the text of which follows: -+.\" * -+.\" * Redistribution and use in source and binary forms, with or without -+.\" * modification, are permitted provided that the following conditions are met: -+.\" * -+.\" * - Redistributions of source code must retain the above copyright notice, -+.\" * this list of conditions and the following disclaimer. -+.\" * - Redistributions in binary form must reproduce the above copyright notice, -+.\" * this list of conditions and the following disclaimer in the documentation -+.\" * and/or other materials provided with the distribution. -+.\" * - Neither the name of the Red Hat, Inc. nor the names of its -+.\" * contributors may be used to endorse or promote products derived from this -+.\" * software without specific prior written permission. -+.\" * -+.\" * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -+.\" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -+.\" * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -+.\" * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -+.\" * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -+.\" * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -+.\" * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -+.\" * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -+.\" * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -+.\" * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF -+.\" * THE POSSIBILITY OF SUCH DAMAGE. -+.\" */ -+.TH "CMAP_KEYS" 7 "2018-10-08" "corosync Man Page" "Corosync Cluster Engine Programmer's Manual" -+ -+.SH NAME -+.P -+cmap_keys \- Overview of keys stored in the Configuration Map -+ -+.SH OVERVIEW -+.P -+There are 3 main types of keys stored in CMAP: -+.PP -+* Mapping of values stored in the config file. -+.PP -+* Runtime statistics. -+.PP -+* Other user created values. -+ -+In this man page, wild-cards have the usual meaning. -+ -+.SH ICMAP KEYS -+These keys are in the icmap (default) map -+.TP -+internal_configuration.* -+Internal configuration data. All keys in this prefix are read only. -+It's only useful for getting a list of loaded services. -+ -+.TP -+logging.* -+Values read from the configuration file. It's possible to change them at runtime. -+If subsystem specific configuration is needed, the key must be in the form -+logging.logger_subsys.SERVICE.key, where SERVICE is upper case name of the service and -+key is same as in the configuration file. All values are of string type. -+ -+.TP -+nodelist.* -+Values are read from the configuration file only (dynamic updates are not allowed). -+Each node element in the configuration file gets -+assigned its position starting from zero. So the first node from the config file has -+nodelist.node.0. prefix. To be a valid entry, each node must have -+.B ring0_addr -+key. -+To change the -+.B nodeid -+key, use a u32 data type. -+ -+Local node position is stored in -+.B local_node_pos -+key (RO), so it's easy to find -+out nodeid/ring addresses of the local node directly from cmap. -+ -+.TP -+runtime.blackbox.* -+Trigger keys for storing fplay data. It's recommended that you use the corosync-blackbox command -+to change keys in this prefix. -+ -+.TP -+runtime.force_gather -+Set to 'yes' to force the processor to move into the GATHER state. This operation -+is dangerous and is not recommended. -+ -+.TP -+runtime.config.* -+Contains the values actually in use by the totem membership protocol. -+Values here are either taken from the Corosync configuration file, -+defaults or computed from entries in the config file. For information -+on individual keys please refer to the man page -+.BR corosync.conf (5). -+ -+.TP -+runtime.services.* -+Prefix with statistics for service engines. Each service has its own -+.B service_id -+key in the prefix with the name runtime.services.SERVICE., where SERVICE is the lower case -+name of the service. Inside the service prefix is the number of messages received and sent -+by the corosync engine in the format runtime.services.SERVICE.EXEC_CALL.rx and -+runtime.services.SERVICE.EXEC_CALL.tx, where EXEC_CALL is the internal id of the service -+call (so for example 3 in cpg service is receive of multicast message from other -+nodes). -+ -+.TP -+runtime.totem.members.* -+Prefix containing members of the totem single ring protocol. Each member -+keys has format runtime.totem.members.NODEID.KEY, where key is -+one of: -+ -+.B config_version -+Config version of the member node. -+ -+.TP -+resources.process.PID.* -+Prefix created by applications using SAM with CMAP integration. -+It contains the following keys: -+ -+.B recovery -+Recovery policy of the process. Can be one of quit or restart. -+ -+.B poll_period -+Value passed in sam_initialize as a time_interval. -+ -+.B last_updated -+Last time SAM received a heartbeat from the client. -+ -+.B state -+State of the client. Can be one of failed, stopped, running and waiting for quorum. -+ -+.TP -+uidgid.* -+Information about users/groups which are allowed to make IPC connections to -+corosync. Entries loaded from configuration file are stored with -+uidgid.config.* prefix and are pruned on configuration file reload. Dynamic -+entries has uidgid.* prefix and a configuration file reload doesn't affect them. -+ -+.TP -+quorum.cancel_wait_for_all -+Tells votequorum to cancel waiting for all nodes at cluster startup. Can be used -+to unblock quorum if notes are known to be down. For pcs use only. -+ -+.TP -+config.reload_in_progress -+This value will be set to 1 (or created) when a corosync.conf reload is started, -+and set to 0 when the reload is completed. This allows interested subsystems -+to do atomic reconfiguration rather than changing each key. Note that -+individual add/change/delete notifications will still be sent during a reload. -+ -+.TP -+config.totemconfig_reload_in_progress -+This key is similar to -+.B config.totemconfig_reload_in_progress -+but changed after the totem config trigger is processed. It is useful (mainly) -+for situations when -+.B nodelist.local_node_pos -+must be correctly reinstated before anything else. -+ -+.SH STATS KEYS -+These keys are in the stats map. All keys in this map are read-only. -+Modification tracking of individual keys is supported in the stats map, but not -+prefixes. Add/Delete operations are supported on prefixes though so you can track -+for new ipc connections or knet interfaces. -+.TP -+stats.srp.* -+Prefix containing statistics about totem. -+Typical key prefixes: -+ -+.B commit_entered -+Number of times the processor entered COMMIT state. -+ -+.B commit_token_lost -+Number of times the processor lost token in COMMIT state. -+ -+.B consensus_timeouts -+How many times the processor timed out forming a consensus about membership. -+ -+.B continuous_gather -+How many times the processor was not able to reach consensus. -+ -+.B firewall_enabled_or_nic_failure -+Set to 1 when processor was not able to reach consensus for long time. The usual -+reason is a badly configured firewall or connection failure. -+ -+.B gather_entered -+Number of times the processor entered GATHER state. -+ -+.B gather_token_lost -+Number of times the processor lost token in GATHER state. -+ -+.B mcast_retx -+Number of retransmitted messages. -+ -+.B mcast_rx -+Number of received multicast messages. -+ -+.B mcast_tx -+Number of transmitted multicast messages. -+ -+.B memb_commit_token_rx -+Number of received commit tokens. -+ -+.B memb_commit_token_tx -+Number of transmitted commit tokens. -+ -+.B memb_join_rx -+Number of received join messages. -+ -+.B memb_join_tx -+Number of transmitted join messages. -+ -+.B memb_merge_detect_rx -+Number of received member merge messages. -+ -+.B memb_merge_detect_tx -+Number of transmitted member merge messages. -+ -+.B orf_token_rx -+Number of received orf tokens. -+ -+.B orf_token_tx -+Number of transmitted orf tokens. -+ -+.B recovery_entered -+Number of times the processor entered recovery. -+ -+.B recovery_token_lost -+Number of times the token was lost in recovery state. -+ -+.B rx_msg_dropped -+Number of received messages which were dropped because they were not expected -+(as example multicast message in commit state). -+ -+.B token_hold_cancel_rx -+Number of received token hold cancel messages. -+ -+.B token_hold_cancel_tx -+Number of transmitted token hold cancel messages. -+ -+.B mtt_rx_token -+Mean transit time of token in milliseconds. In other words, time between -+two consecutive token receives. -+ -+.B avg_token_workload -+Average time in milliseconds of holding time of token on the current processor. -+ -+.B avg_backlog_calc -+Average number of not yet sent messages on the current processor. -+ -+.TP -+stats.knet.nodeX.linkY.* -+Statistics about the network traffic to and from each node and link when using -+tke kronosnet transport -+ -+.B connected -+Whether the link is connected or not -+ -+.B up_count -+Number of times this link has changed state to UP -+ -+.B down_count -+Number of times this link has changed state to DOWN -+ -+.B latency_ave / latency_max / latency_max -+Calculated latencies of this link. Note that if there has been no traffic -+on the link then latency_min will show a very large number. -+ -+.B latency_samples -+The number of samples used to calculate the latency figures, so you have -+some idea of their precision. -+ -+.B rx_data_packets / tx_data_packets -+The number of packets sent/received on this link -+ -+.B rx_data_bytes / tx_data_bytes -+The number of bytes sent/received on this link -+ -+.B rx_pmtu_packets / tx_pmtu_packets -+The number of packets sent/received by the PMTUd subsystem -+ -+.B rx_pmtu_bytes / tx_pmtu_bytes -+The number of bytes sent/received by the PMTUd subsystem -+ -+.B rx_ping_packets / tx_ping_packets -+The number of packets sent/received as pings -+ -+.B rx_ping_bytes / tx_ping_bytes -+The number of bytes sent/received as pings -+ -+.B rx_pong_packets / tx_pong_packets -+The number of packets sent/received as pongs -+ -+.B rx_pong_bytes / tx_pong_bytes -+The number of bytes sent/received as pongs -+ -+.B rx_total_packets / tx_total_packets -+The total number of packets sent/received. The aggregate of all of the above packet stats -+ -+.B rx_total_bytes / tx_total_bytes -+The total number of bytes sent/received. The aggregate of all of the above bytes stats -+ -+.B tx_data_retries / tx_pmtu_retries / tx_ping_retries / tx_pong_retries / tx_total_retries -+Number of times a transmit operation had to be retried due to the socket returning EAGAIN -+ -+.TP -+stats.ipcs.* -+There is information about total number of active connections from client programs -+at the time the request was made. -+.B active -+number of closed connections during whole runtime of corosync -+.B closed -+Total number of connections that have been made since corosync was started -+ -+.TP -+stats.ipcs.ID.* -+Each IPC connection has a unique ID. This is in the form [[serviceX:][PID:]internal_id. -+ -+Typical keys in this prefix are: -+ -+.B proc_name -+process name of connected process (unavailable on some platforms) -+ -+.B dispatched -+number of dispatched messages. -+ -+.B invalid_request -+number of requests made by IPC which are invalid (calling non-existing call, ...). -+ -+.B name -+contains short name of the IPC connection (unavailable on some platforms). -+ -+.B overload -+is number of requests which were not processed because of overload. -+ -+.B queue_size -+contains the number of messages in the queue waiting for send. -+ -+.B recv_retries -+is the total number of interrupted receives. -+ -+.B requests -+contains the number of requests made by IPC. -+ -+.B responses -+is the number of responses sent to the IPC client. -+ -+.B send_retries -+contains the total number of interrupted sends. -+ -+.B service_id -+contains the ID of service which the IPC is connected to. -+ -+.TP -+stats.clear.* -+These are write-only keys used to clear the stats for various subsystems -+ -+.B totem -+Clears the pg & srp totem stats. -+ -+.B knet -+Clears the knet stats -+ -+.B ipc -+Clears the ipc stats -+ -+.B all -+Clears all of the above stats -+ -+ -+.SH DYNAMIC CHANGE USER/GROUP PERMISSION TO USE COROSYNC IPC -+Is the same as in the configuration file. eg: to add UID 500 use -+ -+.br -+# corosync-cmapctl -s uidgid.uid.500 u8 1 -+ -+GID is similar, so to add a GID use -+ -+.br -+# corosync-cmapctl -s uidgid.gid.500 u8 1 -+ -+For removal of permissions, simply delete the key -+ -+.br -+# corosync-cmapctl -d uidgid.gid.500 -+ -+ -+.SH "SEE ALSO" -+.BR corosync_overview (7), -+.BR corosync.conf (5), -+.BR corosync-cmapctl (8) -diff --git a/man/cmap_keys.8 b/man/cmap_keys.8 -deleted file mode 100644 -index e2ea1fb..0000000 ---- a/man/cmap_keys.8 -+++ /dev/null -@@ -1,397 +0,0 @@ --.\"/* --.\" * Copyright (c) 2012-2018 Red Hat, Inc. --.\" * --.\" * All rights reserved. --.\" * --.\" * Author: Jan Friesse (jfriesse@redhat.com) --.\" * --.\" * This software licensed under BSD license, the text of which follows: --.\" * --.\" * Redistribution and use in source and binary forms, with or without --.\" * modification, are permitted provided that the following conditions are met: --.\" * --.\" * - Redistributions of source code must retain the above copyright notice, --.\" * this list of conditions and the following disclaimer. --.\" * - Redistributions in binary form must reproduce the above copyright notice, --.\" * this list of conditions and the following disclaimer in the documentation --.\" * and/or other materials provided with the distribution. --.\" * - Neither the name of the Red Hat, Inc. nor the names of its --.\" * contributors may be used to endorse or promote products derived from this --.\" * software without specific prior written permission. --.\" * --.\" * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" --.\" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE --.\" * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE --.\" * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE --.\" * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR --.\" * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF --.\" * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS --.\" * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN --.\" * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) --.\" * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF --.\" * THE POSSIBILITY OF SUCH DAMAGE. --.\" */ --.TH "CMAP_KEYS" 8 "2018-10-08" "corosync Man Page" "Corosync Cluster Engine Programmer's Manual" -- --.SH NAME --.P --cmap_keys \- Overview of keys stored in the Configuration Map -- --.SH OVERVIEW --.P --There are 3 main types of keys stored in CMAP: --.PP --* Mapping of values stored in the config file. --.PP --* Runtime statistics. --.PP --* Other user created values. -- --In this man page, wild-cards have the usual meaning. -- --.SH ICMAP KEYS --These keys are in the icmap (default) map --.TP --internal_configuration.* --Internal configuration data. All keys in this prefix are read only. --It's only useful for getting a list of loaded services. -- --.TP --logging.* --Values read from the configuration file. It's possible to change them at runtime. --If subsystem specific configuration is needed, the key must be in the form --logging.logger_subsys.SERVICE.key, where SERVICE is upper case name of the service and --key is same as in the configuration file. All values are of string type. -- --.TP --nodelist.* --Values are read from the configuration file only (dynamic updates are not allowed). --Each node element in the configuration file gets --assigned its position starting from zero. So the first node from the config file has --nodelist.node.0. prefix. To be a valid entry, each node must have --.B ring0_addr --key. --To change the --.B nodeid --key, use a u32 data type. -- --Local node position is stored in --.B local_node_pos --key (RO), so it's easy to find --out nodeid/ring addresses of the local node directly from cmap. -- --.TP --runtime.blackbox.* --Trigger keys for storing fplay data. It's recommended that you use the corosync-blackbox command --to change keys in this prefix. -- --.TP --runtime.force_gather --Set to 'yes' to force the processor to move into the GATHER state. This operation --is dangerous and is not recommended. -- --.TP --runtime.config.* --Contains the values actually in use by the totem membership protocol. --Values here are either taken from the Corosync configuration file, --defaults or computed from entries in the config file. For information --on individual keys please refer to the man page --.BR corosync.conf (5). -- --.TP --runtime.services.* --Prefix with statistics for service engines. Each service has its own --.B service_id --key in the prefix with the name runtime.services.SERVICE., where SERVICE is the lower case --name of the service. Inside the service prefix is the number of messages received and sent --by the corosync engine in the format runtime.services.SERVICE.EXEC_CALL.rx and --runtime.services.SERVICE.EXEC_CALL.tx, where EXEC_CALL is the internal id of the service --call (so for example 3 in cpg service is receive of multicast message from other --nodes). -- --.TP --runtime.totem.members.* --Prefix containing members of the totem single ring protocol. Each member --keys has format runtime.totem.members.NODEID.KEY, where key is --one of: -- --.B config_version --Config version of the member node. -- --.TP --resources.process.PID.* --Prefix created by applications using SAM with CMAP integration. --It contains the following keys: -- --.B recovery --Recovery policy of the process. Can be one of quit or restart. -- --.B poll_period --Value passed in sam_initialize as a time_interval. -- --.B last_updated --Last time SAM received a heartbeat from the client. -- --.B state --State of the client. Can be one of failed, stopped, running and waiting for quorum. -- --.TP --uidgid.* --Information about users/groups which are allowed to make IPC connections to --corosync. Entries loaded from configuration file are stored with --uidgid.config.* prefix and are pruned on configuration file reload. Dynamic --entries has uidgid.* prefix and a configuration file reload doesn't affect them. -- --.TP --quorum.cancel_wait_for_all --Tells votequorum to cancel waiting for all nodes at cluster startup. Can be used --to unblock quorum if notes are known to be down. For pcs use only. -- --.TP --config.reload_in_progress --This value will be set to 1 (or created) when a corosync.conf reload is started, --and set to 0 when the reload is completed. This allows interested subsystems --to do atomic reconfiguration rather than changing each key. Note that --individual add/change/delete notifications will still be sent during a reload. -- --.TP --config.totemconfig_reload_in_progress --This key is similar to --.B config.totemconfig_reload_in_progress --but changed after the totem config trigger is processed. It is useful (mainly) --for situations when --.B nodelist.local_node_pos --must be correctly reinstated before anything else. -- --.SH STATS KEYS --These keys are in the stats map. All keys in this map are read-only. --Modification tracking of individual keys is supported in the stats map, but not --prefixes. Add/Delete operations are supported on prefixes though so you can track --for new ipc connections or knet interfaces. --.TP --stats.srp.* --Prefix containing statistics about totem. --Typical key prefixes: -- --.B commit_entered --Number of times the processor entered COMMIT state. -- --.B commit_token_lost --Number of times the processor lost token in COMMIT state. -- --.B consensus_timeouts --How many times the processor timed out forming a consensus about membership. -- --.B continuous_gather --How many times the processor was not able to reach consensus. -- --.B firewall_enabled_or_nic_failure --Set to 1 when processor was not able to reach consensus for long time. The usual --reason is a badly configured firewall or connection failure. -- --.B gather_entered --Number of times the processor entered GATHER state. -- --.B gather_token_lost --Number of times the processor lost token in GATHER state. -- --.B mcast_retx --Number of retransmitted messages. -- --.B mcast_rx --Number of received multicast messages. -- --.B mcast_tx --Number of transmitted multicast messages. -- --.B memb_commit_token_rx --Number of received commit tokens. -- --.B memb_commit_token_tx --Number of transmitted commit tokens. -- --.B memb_join_rx --Number of received join messages. -- --.B memb_join_tx --Number of transmitted join messages. -- --.B memb_merge_detect_rx --Number of received member merge messages. -- --.B memb_merge_detect_tx --Number of transmitted member merge messages. -- --.B orf_token_rx --Number of received orf tokens. -- --.B orf_token_tx --Number of transmitted orf tokens. -- --.B recovery_entered --Number of times the processor entered recovery. -- --.B recovery_token_lost --Number of times the token was lost in recovery state. -- --.B rx_msg_dropped --Number of received messages which were dropped because they were not expected --(as example multicast message in commit state). -- --.B token_hold_cancel_rx --Number of received token hold cancel messages. -- --.B token_hold_cancel_tx --Number of transmitted token hold cancel messages. -- --.B mtt_rx_token --Mean transit time of token in milliseconds. In other words, time between --two consecutive token receives. -- --.B avg_token_workload --Average time in milliseconds of holding time of token on the current processor. -- --.B avg_backlog_calc --Average number of not yet sent messages on the current processor. -- --.TP --stats.knet.nodeX.linkY.* --Statistics about the network traffic to and from each node and link when using --tke kronosnet transport -- --.B connected --Whether the link is connected or not -- --.B up_count --Number of times this link has changed state to UP -- --.B down_count --Number of times this link has changed state to DOWN -- --.B latency_ave / latency_max / latency_max --Calculated latencies of this link. Note that if there has been no traffic --on the link then latency_min will show a very large number. -- --.B latency_samples --The number of samples used to calculate the latency figures, so you have --some idea of their precision. -- --.B rx_data_packets / tx_data_packets --The number of packets sent/received on this link -- --.B rx_data_bytes / tx_data_bytes --The number of bytes sent/received on this link -- --.B rx_pmtu_packets / tx_pmtu_packets --The number of packets sent/received by the PMTUd subsystem -- --.B rx_pmtu_bytes / tx_pmtu_bytes --The number of bytes sent/received by the PMTUd subsystem -- --.B rx_ping_packets / tx_ping_packets --The number of packets sent/received as pings -- --.B rx_ping_bytes / tx_ping_bytes --The number of bytes sent/received as pings -- --.B rx_pong_packets / tx_pong_packets --The number of packets sent/received as pongs -- --.B rx_pong_bytes / tx_pong_bytes --The number of bytes sent/received as pongs -- --.B rx_total_packets / tx_total_packets --The total number of packets sent/received. The aggregate of all of the above packet stats -- --.B rx_total_bytes / tx_total_bytes --The total number of bytes sent/received. The aggregate of all of the above bytes stats -- --.B tx_data_retries / tx_pmtu_retries / tx_ping_retries / tx_pong_retries / tx_total_retries --Number of times a transmit operation had to be retried due to the socket returning EAGAIN -- --.TP --stats.ipcs.* --There is information about total number of active connections from client programs --at the time the request was made. --.B active --number of closed connections during whole runtime of corosync --.B closed --Total number of connections that have been made since corosync was started -- --.TP --stats.ipcs.ID.* --Each IPC connection has a unique ID. This is in the form [[serviceX:][PID:]internal_id. -- --Typical keys in this prefix are: -- --.B proc_name --process name of connected process (unavailable on some platforms) -- --.B dispatched --number of dispatched messages. -- --.B invalid_request --number of requests made by IPC which are invalid (calling non-existing call, ...). -- --.B name --contains short name of the IPC connection (unavailable on some platforms). -- --.B overload --is number of requests which were not processed because of overload. -- --.B queue_size --contains the number of messages in the queue waiting for send. -- --.B recv_retries --is the total number of interrupted receives. -- --.B requests --contains the number of requests made by IPC. -- --.B responses --is the number of responses sent to the IPC client. -- --.B send_retries --contains the total number of interrupted sends. -- --.B service_id --contains the ID of service which the IPC is connected to. -- --.TP --stats.clear.* --These are write-only keys used to clear the stats for various subsystems -- --.B totem --Clears the pg & srp totem stats. -- --.B knet --Clears the knet stats -- --.B ipc --Clears the ipc stats -- --.B all --Clears all of the above stats -- -- --.SH DYNAMIC CHANGE USER/GROUP PERMISSION TO USE COROSYNC IPC --Is the same as in the configuration file. eg: to add UID 500 use -- --.br --# corosync-cmapctl -s uidgid.uid.500 u8 1 -- --GID is similar, so to add a GID use -- --.br --# corosync-cmapctl -s uidgid.gid.500 u8 1 -- --For removal of permissions, simply delete the key -- --.br --# corosync-cmapctl -d uidgid.gid.500 -- -- --.SH "SEE ALSO" --.BR corosync_overview (7), --.BR corosync.conf (5), --.BR corosync-cmapctl (8) -diff --git a/man/cmap_overview.3 b/man/cmap_overview.3 -index cf4cabb..0aa3c14 100644 ---- a/man/cmap_overview.3 -+++ b/man/cmap_overview.3 -@@ -54,7 +54,7 @@ The library provides a mechanism to: - .PP - * Track changes on keys - --Description of most keys created by corosync itself can be found in cmap_keys (8). -+Description of most keys created by corosync itself can be found in cmap_keys (7). - - .SH BUGS - .SH "SEE ALSO" -@@ -75,4 +75,4 @@ Description of most keys created by corosync itself can be found in cmap_keys (8 - .BR cmap_iter_finalize (3), - .BR cmap_track_add (3), - .BR cmap_track_delete (3), --.BR cmap_keys (8) -+.BR cmap_keys (7) -diff --git a/man/corosync-cmapctl.8 b/man/corosync-cmapctl.8 -index 637e597..8826503 100644 ---- a/man/corosync-cmapctl.8 -+++ b/man/corosync-cmapctl.8 -@@ -96,4 +96,4 @@ corosync\-cmapctl \fB\-C\fR [ipc|totem|knet|all] - - .SH "SEE ALSO" - .BR cmap_overview (3), --.BR cmap_keys (8) -+.BR cmap_keys (7) -diff --git a/man/index.html b/man/index.html -index f4819e5..21326dc 100644 ---- a/man/index.html -+++ b/man/index.html -@@ -63,7 +63,7 @@ - Description of corosync-cmapctl tool. -
- -- cmap_keys(8): -+ cmap_keys(7): - Overview of keys stored in the Configuration Map. -
- --- -1.8.3.1 - diff --git a/SOURCES/bz1791792-2-stats-Add-stats-for-scheduler-misses.patch b/SOURCES/bz1791792-2-stats-Add-stats-for-scheduler-misses.patch deleted file mode 100644 index 898d9b5..0000000 --- a/SOURCES/bz1791792-2-stats-Add-stats-for-scheduler-misses.patch +++ /dev/null @@ -1,320 +0,0 @@ -From 48b6894ef41e9a06ccbb696d062d86ef60dc2c4b Mon Sep 17 00:00:00 2001 -From: Christine Caulfield -Date: Fri, 17 Jan 2020 14:22:16 +0000 -Subject: [PATCH] stats: Add stats for scheduler misses - -This patch add a stats.schedmiss.* set of entries that -are a record of the last 10 times corosync was not scheduled -in time. - -These entries are keypt in reverse order (so stats.schedmiss.0.* is -always the latest one kept) and the values, including the timestamp, -are in milliseconds. - -It's also possible to use a cmap tracker to follow these events, which -might be useful. - -Signed-off-by: Christine Caulfield -Reviewed-by: Jan Friesse ---- - exec/main.c | 2 + - exec/stats.c | 113 +++++++++++++++++++++++++++++++++++++++++++---- - exec/stats.h | 2 + - man/cmap_keys.7 | 26 ++++++++++- - tools/corosync-cmapctl.c | 5 ++- - 5 files changed, 136 insertions(+), 12 deletions(-) - -diff --git a/exec/main.c b/exec/main.c -index 7a471a1..fb0486e 100644 ---- a/exec/main.c -+++ b/exec/main.c -@@ -835,6 +835,8 @@ static void timer_function_scheduler_timeout (void *data) - log_printf (LOGSYS_LEVEL_WARNING, "Corosync main process was not scheduled for %0.4f ms " - "(threshold is %0.4f ms). Consider token timeout increase.", - (float)tv_diff / QB_TIME_NS_IN_MSEC, (float)timeout_data->max_tv_diff / QB_TIME_NS_IN_MSEC); -+ -+ stats_add_schedmiss_event(tv_current / 1000, (float)tv_diff / QB_TIME_NS_IN_MSEC); - } - - /* -diff --git a/exec/stats.c b/exec/stats.c -index e89504e..d5c1cbc 100644 ---- a/exec/stats.c -+++ b/exec/stats.c -@@ -1,5 +1,5 @@ - /* -- * Copyright (c) 2017 Red Hat, Inc. -+ * Copyright (c) 2017-2020 Red Hat, Inc. - * - * All rights reserved. - * -@@ -60,9 +60,20 @@ LOGSYS_DECLARE_SUBSYS ("STATS"); - - static qb_map_t *stats_map; - -+/* Structure of an element in the schedmiss array */ -+struct schedmiss_entry { -+ uint64_t timestamp; -+ float delay; -+}; -+#define MAX_SCHEDMISS_EVENTS 10 -+static struct schedmiss_entry schedmiss_event[MAX_SCHEDMISS_EVENTS]; -+static unsigned int highest_schedmiss_event; -+ -+#define SCHEDMISS_PREFIX "stats.schedmiss" -+ - /* Convert iterator number to text and a stats pointer */ - struct cs_stats_conv { -- enum {STAT_PG, STAT_SRP, STAT_KNET, STAT_KNET_HANDLE, STAT_IPCSC, STAT_IPCSG} type; -+ enum {STAT_PG, STAT_SRP, STAT_KNET, STAT_KNET_HANDLE, STAT_IPCSC, STAT_IPCSG, STAT_SCHEDMISS} type; - const char *name; - const size_t offset; - const icmap_value_types_t value_type; -@@ -190,6 +201,10 @@ struct cs_stats_conv cs_ipcs_global_stats[] = { - { STAT_IPCSG, "global.active", offsetof(struct ipcs_global_stats, active), ICMAP_VALUETYPE_UINT64}, - { STAT_IPCSG, "global.closed", offsetof(struct ipcs_global_stats, closed), ICMAP_VALUETYPE_UINT64}, - }; -+struct cs_stats_conv cs_schedmiss_stats[] = { -+ { STAT_SCHEDMISS, "timestamp", offsetof(struct schedmiss_entry, timestamp), ICMAP_VALUETYPE_UINT64}, -+ { STAT_SCHEDMISS, "delay", offsetof(struct schedmiss_entry, delay), ICMAP_VALUETYPE_FLOAT}, -+}; - - #define NUM_PG_STATS (sizeof(cs_pg_stats) / sizeof(struct cs_stats_conv)) - #define NUM_SRP_STATS (sizeof(cs_srp_stats) / sizeof(struct cs_stats_conv)) -@@ -286,7 +301,7 @@ cs_error_t stats_map_init(const struct corosync_api_v1 *corosync_api) - stats_add_entry(param, &cs_ipcs_global_stats[i]); - } - -- /* KNET and IPCS stats are added when appropriate */ -+ /* KNET, IPCS & SCHEDMISS stats are added when appropriate */ - return CS_OK; - } - -@@ -307,6 +322,8 @@ cs_error_t stats_map_get(const char *key_name, - int link_no; - int service_id; - uint32_t pid; -+ unsigned int sm_event; -+ char *sm_type; - void *conn_ptr; - - item = qb_map_get(stats_map, key_name); -@@ -363,17 +380,85 @@ cs_error_t stats_map_get(const char *key_name, - cs_ipcs_get_global_stats(&ipcs_global_stats); - stats_map_set_value(statinfo, &ipcs_global_stats, value, value_len, type); - break; -+ case STAT_SCHEDMISS: -+ if (sscanf(key_name, SCHEDMISS_PREFIX ".%d", &sm_event) != 1) { -+ return CS_ERR_NOT_EXIST; -+ } -+ -+ sm_type = strrchr(key_name, '.'); -+ if (sm_type == NULL) { -+ return CS_ERR_NOT_EXIST; -+ } -+ sm_type++; -+ -+ if (strcmp(sm_type, "timestamp") == 0) { -+ memcpy(value, &schedmiss_event[sm_event].timestamp, sizeof(uint64_t)); -+ *value_len = sizeof(uint64_t); -+ *type = ICMAP_VALUETYPE_UINT64; -+ } -+ if (strcmp(sm_type, "delay") == 0) { -+ memcpy(value, &schedmiss_event[sm_event].delay, sizeof(float)); -+ *value_len = sizeof(float); -+ *type = ICMAP_VALUETYPE_FLOAT; -+ } -+ break; - default: - return CS_ERR_LIBRARY; - } - return CS_OK; - } - --#define STATS_CLEAR "stats.clear." --#define STATS_CLEAR_KNET "stats.clear.knet" --#define STATS_CLEAR_IPC "stats.clear.ipc" --#define STATS_CLEAR_TOTEM "stats.clear.totem" --#define STATS_CLEAR_ALL "stats.clear.all" -+static void schedmiss_clear_stats(void) -+{ -+ int i; -+ char param[ICMAP_KEYNAME_MAXLEN]; -+ -+ for (i=0; i=0; i--) { -+ schedmiss_event[i+1].timestamp = schedmiss_event[i].timestamp; -+ schedmiss_event[i+1].delay = schedmiss_event[i].delay; -+ } -+ -+ /* New entries are always at the front */ -+ schedmiss_event[0].timestamp = timestamp; -+ schedmiss_event[0].delay = delay; -+ -+ /* If we've not run off the end then add an entry in the trie for the new 'end' one */ -+ if (highest_schedmiss_event < MAX_SCHEDMISS_EVENTS) { -+ sprintf(param, SCHEDMISS_PREFIX ".%i.timestamp", highest_schedmiss_event); -+ stats_add_entry(param, &cs_schedmiss_stats[0]); -+ sprintf(param, SCHEDMISS_PREFIX ".%i.delay", highest_schedmiss_event); -+ stats_add_entry(param, &cs_schedmiss_stats[1]); -+ highest_schedmiss_event++; -+ } -+ /* Notifications get sent by the stats_updater */ -+} -+ -+#define STATS_CLEAR "stats.clear." -+#define STATS_CLEAR_KNET "stats.clear.knet" -+#define STATS_CLEAR_IPC "stats.clear.ipc" -+#define STATS_CLEAR_TOTEM "stats.clear.totem" -+#define STATS_CLEAR_ALL "stats.clear.all" -+#define STATS_CLEAR_SCHEDMISS "stats.clear.schedmiss" - - cs_error_t stats_map_set(const char *key_name, - const void *value, -@@ -394,9 +479,14 @@ cs_error_t stats_map_set(const char *key_name, - totempg_stats_clear(TOTEMPG_STATS_CLEAR_TOTEM); - cleared = 1; - } -+ if (strncmp(key_name, STATS_CLEAR_SCHEDMISS, strlen(STATS_CLEAR_SCHEDMISS)) == 0) { -+ schedmiss_clear_stats(); -+ cleared = 1; -+ } - if (strncmp(key_name, STATS_CLEAR_ALL, strlen(STATS_CLEAR_ALL)) == 0) { - totempg_stats_clear(TOTEMPG_STATS_CLEAR_TRANSPORT | TOTEMPG_STATS_CLEAR_TOTEM); - cs_ipcs_clear_stats(); -+ schedmiss_clear_stats(); - cleared = 1; - } - if (!cleared) { -@@ -500,6 +590,11 @@ static void stats_map_notify_fn(uint32_t event, char *key, void *old_value, void - return ; - } - -+ /* Ignore schedmiss trackers as the values are read from the circular buffer */ -+ if (strncmp(key, SCHEDMISS_PREFIX, strlen(SCHEDMISS_PREFIX)) == 0 ) { -+ return ; -+ } -+ - new_val.data = new_value; - if (stats_map_get(key, - &new_value, -@@ -556,7 +651,7 @@ cs_error_t stats_map_track_add(const char *key_name, - } - /* Get initial value */ - if (stats_map_get(tracker->key_name, -- &tracker->old_value, &value_len, &type) == CS_OK) { -+ &tracker->old_value, &value_len, &type) != CS_OK) { - tracker->old_value = 0ULL; - } - } else { -diff --git a/exec/stats.h b/exec/stats.h -index 45891ae..eac9e7c 100644 ---- a/exec/stats.h -+++ b/exec/stats.h -@@ -69,3 +69,5 @@ void stats_trigger_trackers(void); - void stats_ipcs_add_connection(int service_id, uint32_t pid, void *ptr); - void stats_ipcs_del_connection(int service_id, uint32_t pid, void *ptr); - cs_error_t cs_ipcs_get_conn_stats(int service_id, uint32_t pid, void *conn_ptr, struct ipcs_conn_stats *ipcs_stats); -+ -+void stats_add_schedmiss_event(uint64_t, float delay); -diff --git a/man/cmap_keys.7 b/man/cmap_keys.7 -index 6bc04fe..da95c51 100644 ---- a/man/cmap_keys.7 -+++ b/man/cmap_keys.7 -@@ -1,5 +1,5 @@ - .\"/* --.\" * Copyright (c) 2012-2018 Red Hat, Inc. -+.\" * Copyright (c) 2012-2020 Red Hat, Inc. - .\" * - .\" * All rights reserved. - .\" * -@@ -357,6 +357,27 @@ contains the total number of interrupted sends. - .B service_id - contains the ID of service which the IPC is connected to. - -+ -+.TP -+stats.schedmiss..* -+If corosync is not scheduled after the required period of time it will -+log this event and also write an entry to the stats cmap under this key. -+There can be up to 10 entries (0..9) in here, when an 11th event happens -+the earliest will be removed. -+ -+These events will always be in reverse order, so stats.schedmiss.0.* will -+always be the latest event kept and 9 the oldest. If you want to listen -+for notifications then you are recommended to listen for changes -+to stats.schedmiss.0.timestamp or stats.schedmiss.0.delay. -+ -+.B timestamp -+The time of the event in ms since the Epoch (ie time_t * 1000 but with -+valid milliseconds). -+ -+.B delay -+The time that corosync was paused (in ms, float value). -+ -+ - .TP - stats.clear.* - These are write-only keys used to clear the stats for various subsystems -@@ -370,6 +391,9 @@ Clears the knet stats - .B ipc - Clears the ipc stats - -+.B schedmiss -+Clears the schedmiss stats -+ - .B all - Clears all of the above stats - -diff --git a/tools/corosync-cmapctl.c b/tools/corosync-cmapctl.c -index a4b61bd..ffca7e1 100644 ---- a/tools/corosync-cmapctl.c -+++ b/tools/corosync-cmapctl.c -@@ -115,7 +115,7 @@ static int print_help(void) - printf(" about the networking and IPC traffic in some detail.\n"); - printf("\n"); - printf("Clear stats:\n"); -- printf(" corosync-cmapctl -C [knet|ipc|totem|all]\n"); -+ printf(" corosync-cmapctl -C [knet|ipc|totem|schedmiss|all]\n"); - printf(" The 'stats' map is implied\n"); - printf("\n"); - printf("Load settings from a file:\n"); -@@ -849,6 +849,7 @@ int main(int argc, char *argv[]) - if (strcmp(optarg, "knet") == 0 || - strcmp(optarg, "totem") == 0 || - strcmp(optarg, "ipc") == 0 || -+ strcmp(optarg, "schedmiss") == 0 || - strcmp(optarg, "all") == 0) { - action = ACTION_CLEARSTATS; - clear_opt = optarg; -@@ -857,7 +858,7 @@ int main(int argc, char *argv[]) - map = CMAP_MAP_STATS; - } - else { -- fprintf(stderr, "argument to -C should be 'knet', 'totem', 'ipc' or 'all'\n"); -+ fprintf(stderr, "argument to -C should be 'knet', 'totem', 'ipc', 'schedmiss' or 'all'\n"); - return (EXIT_FAILURE); - } - break; --- -1.8.3.1 - diff --git a/SOURCES/bz1791792-3-stats-Use-nanoseconds-from-epoch-for-schedmiss.patch b/SOURCES/bz1791792-3-stats-Use-nanoseconds-from-epoch-for-schedmiss.patch deleted file mode 100644 index 8f6f5f1..0000000 --- a/SOURCES/bz1791792-3-stats-Use-nanoseconds-from-epoch-for-schedmiss.patch +++ /dev/null @@ -1,31 +0,0 @@ -From ebd05fa00826c366922e619b012a0684c6856539 Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Thu, 23 Jan 2020 17:11:54 +0100 -Subject: [PATCH] stats: Use nanoseconds from epoch for schedmiss - -Using monotonic time is not working because it doesn't have to match -time from epoch. - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - exec/main.c | 3 ++- - 1 file changed, 2 insertions(+), 1 deletion(-) - -diff --git a/exec/main.c b/exec/main.c -index fb0486e..821d97e 100644 ---- a/exec/main.c -+++ b/exec/main.c -@@ -836,7 +836,8 @@ static void timer_function_scheduler_timeout (void *data) - "(threshold is %0.4f ms). Consider token timeout increase.", - (float)tv_diff / QB_TIME_NS_IN_MSEC, (float)timeout_data->max_tv_diff / QB_TIME_NS_IN_MSEC); - -- stats_add_schedmiss_event(tv_current / 1000, (float)tv_diff / QB_TIME_NS_IN_MSEC); -+ stats_add_schedmiss_event(qb_util_nano_from_epoch_get() / QB_TIME_NS_IN_MSEC, -+ (float)tv_diff / QB_TIME_NS_IN_MSEC); - } - - /* --- -1.8.3.1 - diff --git a/SOURCES/bz1791792-4-main-Add-schedmiss-timestamp-into-message.patch b/SOURCES/bz1791792-4-main-Add-schedmiss-timestamp-into-message.patch deleted file mode 100644 index 44803e2..0000000 --- a/SOURCES/bz1791792-4-main-Add-schedmiss-timestamp-into-message.patch +++ /dev/null @@ -1,47 +0,0 @@ -From 35662dd0ec53f456445c30c0ef92892f47b25aa2 Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Mon, 24 Feb 2020 14:58:45 +0100 -Subject: [PATCH] main: Add schedmiss timestamp into message - -This is useful for matching schedmiss event in stats map with logged -event. - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - exec/main.c | 9 ++++++--- - 1 file changed, 6 insertions(+), 3 deletions(-) - -diff --git a/exec/main.c b/exec/main.c -index 821d97e..8c3df79 100644 ---- a/exec/main.c -+++ b/exec/main.c -@@ -817,6 +817,7 @@ static void timer_function_scheduler_timeout (void *data) - struct scheduler_pause_timeout_data *timeout_data = (struct scheduler_pause_timeout_data *)data; - unsigned long long tv_current; - unsigned long long tv_diff; -+ uint64_t schedmiss_event_tstamp; - - tv_current = qb_util_nano_current_get (); - -@@ -832,12 +833,14 @@ static void timer_function_scheduler_timeout (void *data) - timeout_data->tv_prev = tv_current; - - if (tv_diff > timeout_data->max_tv_diff) { -- log_printf (LOGSYS_LEVEL_WARNING, "Corosync main process was not scheduled for %0.4f ms " -+ schedmiss_event_tstamp = qb_util_nano_from_epoch_get() / QB_TIME_NS_IN_MSEC; -+ -+ log_printf (LOGSYS_LEVEL_WARNING, "Corosync main process was not scheduled (@%" PRIu64 ") for %0.4f ms " - "(threshold is %0.4f ms). Consider token timeout increase.", -+ schedmiss_event_tstamp, - (float)tv_diff / QB_TIME_NS_IN_MSEC, (float)timeout_data->max_tv_diff / QB_TIME_NS_IN_MSEC); - -- stats_add_schedmiss_event(qb_util_nano_from_epoch_get() / QB_TIME_NS_IN_MSEC, -- (float)tv_diff / QB_TIME_NS_IN_MSEC); -+ stats_add_schedmiss_event(schedmiss_event_tstamp, (float)tv_diff / QB_TIME_NS_IN_MSEC); - } - - /* --- -1.8.3.1 - diff --git a/SOURCES/bz1809864-1-votequorum-Change-check-of-expected_votes.patch b/SOURCES/bz1809864-1-votequorum-Change-check-of-expected_votes.patch deleted file mode 100644 index 06d444e..0000000 --- a/SOURCES/bz1809864-1-votequorum-Change-check-of-expected_votes.patch +++ /dev/null @@ -1,51 +0,0 @@ -From 0c16442f2d93f32a229b87d2672e2dc8025ec704 Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Wed, 4 Mar 2020 11:42:15 +0100 -Subject: [PATCH] votequorum: Change check of expected_votes - -Previously value of new expected_votes was checked so newly computed -quorum value was in the interval . The -upper range prevented the cluster to become unquorate, but bottom check -was almost useless because it allowed to change expected_votes so it is -smaller than total_votes. - -Solution is to check if expected_votes is bigger or equal to total_votes -and for quorate cluster only check if cluster doesn't become unquorate -(for unquorate cluster one can set upper range freely - as it is -perfectly possible when using config file) - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - exec/votequorum.c | 10 +++++++--- - 1 file changed, 7 insertions(+), 3 deletions(-) - -diff --git a/exec/votequorum.c b/exec/votequorum.c -index 52424fa..b152425 100644 ---- a/exec/votequorum.c -+++ b/exec/votequorum.c -@@ -1,5 +1,5 @@ - /* -- * Copyright (c) 2009-2015 Red Hat, Inc. -+ * Copyright (c) 2009-2020 Red Hat, Inc. - * - * All rights reserved. - * -@@ -2688,8 +2688,12 @@ static void message_handler_req_lib_votequorum_setexpected (void *conn, const vo - */ - newquorum = calculate_quorum(1, req_lib_votequorum_setexpected->expected_votes, &total_votes); - allow_downscale = allow_downscale_status; -- if (newquorum < total_votes / 2 || -- newquorum > total_votes) { -+ /* -+ * Setting expected_votes < total_votes doesn't make sense. -+ * For quorate cluster prevent cluster to become unquorate. -+ */ -+ if (req_lib_votequorum_setexpected->expected_votes < total_votes || -+ (cluster_is_quorate && (newquorum > total_votes))) { - error = CS_ERR_INVALID_PARAM; - goto error_exit; - } --- -1.8.3.1 - diff --git a/SOURCES/bz1809864-2-quorumtool-exit-on-invalid-expected-votes.patch b/SOURCES/bz1809864-2-quorumtool-exit-on-invalid-expected-votes.patch deleted file mode 100644 index 1f4c760..0000000 --- a/SOURCES/bz1809864-2-quorumtool-exit-on-invalid-expected-votes.patch +++ /dev/null @@ -1,33 +0,0 @@ -From 5f543465bb3506b7f4929a426f1c22a9c854cecd Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Wed, 4 Mar 2020 08:53:41 +0100 -Subject: [PATCH] quorumtool: exit on invalid expected votes - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - tools/corosync-quorumtool.c | 3 ++- - 1 file changed, 2 insertions(+), 1 deletion(-) - -diff --git a/tools/corosync-quorumtool.c b/tools/corosync-quorumtool.c -index 9bef844..44bf181 100644 ---- a/tools/corosync-quorumtool.c -+++ b/tools/corosync-quorumtool.c -@@ -1,5 +1,5 @@ - /* -- * Copyright (c) 2009-2019 Red Hat, Inc. -+ * Copyright (c) 2009-2020 Red Hat, Inc. - * - * All rights reserved. - * -@@ -937,6 +937,7 @@ int main (int argc, char *argv[]) { - votes = strtol(optarg, &endptr, 0); - if ((votes == 0 && endptr == optarg) || votes <= 0) { - fprintf(stderr, "New expected votes value was not valid, try a positive number\n"); -+ exit(EXIT_FAILURE); - } else { - command_opt = CMD_SETEXPECTED; - } --- -1.8.3.1 - diff --git a/SOURCES/bz1816653-1-votequorum-set-wfa-status-only-on-startup.patch b/SOURCES/bz1816653-1-votequorum-set-wfa-status-only-on-startup.patch deleted file mode 100644 index 367f44c..0000000 --- a/SOURCES/bz1816653-1-votequorum-set-wfa-status-only-on-startup.patch +++ /dev/null @@ -1,67 +0,0 @@ -From ca320beac25f82c0c555799e647a47975a333c28 Mon Sep 17 00:00:00 2001 -From: Jan Friesse -Date: Tue, 10 Mar 2020 17:49:27 +0100 -Subject: [PATCH] votequorum: set wfa status only on startup - -Previously reload of configuration with enabled wait_for_all result in -set of wait_for_all_status which set cluster_is_quorate to 0 but didn't -inform the quorum service so votequorum and quorum information may get -out of sync. - -Example is 1 node cluster, which is extended to 3 nodes. Quorum service -reports cluster as a quorate (incorrect) and votequorum as not-quorate -(correct). Similar behavior happens when extending cluster in general, -but some configurations are less incorrect (3->4). - -Discussed solution was to inform quorum service but that would mean -every reload would cause loss of quorum until all nodes would be seen -again. - -Such behaviour is consistent but seems to be a bit too strict. - -Proposed solution sets wait_for_all_status only on startup and -doesn't touch it during reload. - -This solution fulfills requirement of "cluster will be quorate for -the first time only after all nodes have been visible at least -once at the same time." because node clears wait_for_all_status only -after it sees all other nodes or joins cluster which is quorate. It also -solves problem with extending cluster, because when cluster becomes -unquorate (1->3) wait_for_all_status is set. - -Added assert is only for ensure that I haven't missed any case when -quorate cluster may become unquorate. - -Signed-off-by: Jan Friesse -Reviewed-by: Christine Caulfield ---- - exec/votequorum.c | 6 ++++-- - 1 file changed, 4 insertions(+), 2 deletions(-) - -diff --git a/exec/votequorum.c b/exec/votequorum.c -index b152425..fb9f1cd 100644 ---- a/exec/votequorum.c -+++ b/exec/votequorum.c -@@ -1009,7 +1009,7 @@ static void are_we_quorate(unsigned int total_votes) - "Waiting for all cluster members. " - "Current votes: %d expected_votes: %d", - total_votes, us->expected_votes); -- cluster_is_quorate = 0; -+ assert(!cluster_is_quorate); - return; - } - update_wait_for_all_status(0); -@@ -1547,7 +1547,9 @@ static char *votequorum_readconfig(int runtime) - update_ev_barrier(us->expected_votes); - update_two_node(); - if (wait_for_all) { -- update_wait_for_all_status(1); -+ if (!runtime) { -+ update_wait_for_all_status(1); -+ } - } else if (wait_for_all_autoset && wait_for_all_status) { - /* - * Reset wait for all status for consistency when wfa is auto-unset by 2node. --- -1.8.3.1 - diff --git a/SPECS/corosync.spec b/SPECS/corosync.spec index 17800d5..4e16761 100644 --- a/SPECS/corosync.spec +++ b/SPECS/corosync.spec @@ -22,22 +22,12 @@ Name: corosync Summary: The Corosync Cluster Engine and Application Programming Interfaces -Version: 3.0.3 -Release: 4%{?gitver}%{?dist} +Version: 3.1.0 +Release: 1%{?gitver}%{?dist} License: BSD URL: http://corosync.github.io/corosync/ Source0: http://build.clusterlabs.org/corosync/releases/%{name}-%{version}%{?gittarver}.tar.gz -Patch0: bz1780137-1-votequorum-Ignore-the-icmap_get_-return-value.patch -Patch1: bz1791792-1-man-move-cmap_keys-man-page-from-section-8-to-7.patch -Patch2: bz1780137-2-votequorum-Reflect-runtime-change-of-2Node-to-WFA.patch -Patch3: bz1791792-2-stats-Add-stats-for-scheduler-misses.patch -Patch4: bz1791792-3-stats-Use-nanoseconds-from-epoch-for-schedmiss.patch -Patch5: bz1791792-4-main-Add-schedmiss-timestamp-into-message.patch -Patch6: bz1809864-1-votequorum-Change-check-of-expected_votes.patch -Patch7: bz1809864-2-quorumtool-exit-on-invalid-expected-votes.patch -Patch8: bz1816653-1-votequorum-set-wfa-status-only-on-startup.patch - %if %{with spausedd} Source1: https://github.com/jfriesse/spausedd/releases/download/%{spausedd_version}/spausedd-%{spausedd_version}.tar.gz # VMGuestLib exists only for x86_64 architecture @@ -50,14 +40,16 @@ Source1: https://github.com/jfriesse/spausedd/releases/download/%{spausedd_versi # The automatic dependency overridden in favor of explicit version lock Requires: corosynclib%{?_isa} = %{version}-%{release} +# Support crypto reload +Requires: libknet1 >= 1.18 # NSS crypto plugin should be always installed -Requires: libknet1-crypto-nss-plugin +Requires: libknet1-crypto-nss-plugin >= 1.18 # Build bits BuildRequires: gcc BuildRequires: groff BuildRequires: libqb-devel -BuildRequires: libknet1-devel +BuildRequires: libknet1-devel >= 1.18 BuildRequires: zlib-devel %if %{with runautogen} BuildRequires: autoconf automake libtool @@ -99,16 +91,6 @@ BuildRequires: pkgconfig(vmguestlib) %setup -q -n %{name}-%{version}%{?gittarver} %endif -%patch0 -p1 -b .bz1780137-1 -%patch1 -p1 -b .bz1791792-1 -%patch2 -p1 -b .bz1780137-2 -%patch3 -p1 -b .bz1791792-2 -%patch4 -p1 -b .bz1791792-3 -%patch5 -p1 -b .bz1791792-4 -%patch6 -p1 -b .bz1809864-1 -%patch7 -p1 -b .bz1809864-2 -%patch8 -p1 -b .bz1816653-1 - %build %if %{with runautogen} ./autogen.sh @@ -407,6 +389,17 @@ fi %endif %changelog +* Tue Oct 20 2020 Jan Friesse 3.1.0-1 +- Resolves: rhbz#1855293 +- Resolves: rhbz#1855303 +- Resolves: rhbz#1870449 +- Resolves: rhbz#1887400 + +- New upstream release (rhbz#1855293) +- Support for reload of crypto configuration (rhbz#1855303) +- Increase default token timeout to 3000ms (rhbz#1870449) +- Add support for nodelist callback into quorum service (rhbz#1887400) + * Tue May 26 2020 Jan Friesse 3.0.3-4 - Resolves: rhbz#1780137 - Resolves: rhbz#1791792