Blob Blame History Raw
From 7d62105c676fc79e0c32766c41cd034655a524ff Mon Sep 17 00:00:00 2001
From: "Endi S. Dewata" <edewata@redhat.com>
Date: Tue, 25 Jan 2022 16:29:53 -0600
Subject: [PATCH] Fix pki-healthcheck for clones

Previously the ClonesConnectivyAndDataCheck.check_kra_clones()
was trying to check KRA clone status by retrieving a key using
the subsystem cert. This operation did not work since the user
associated with the cert did not have access to the keys. The
code has been changed to get the status from GetStatus service
instead. The original code might be moved into IPA later so it
could run with IPA's RA agent credentials which would allow
access to the keys.

Previously the ClonesPlugin.contact_subsystem_using_sslget()
used sslget to call GetStatus service and returned the entire
output which was then incorrectly processed in XML format. The
method has been renamed to get_status() and changed to use
PKIConnection and process the response in either JSON or XML
format, then only return the subsystem status. All callers
have been updated accordingly.

The ClonesPlugin.contact_subsystem_using_pki() is no longer
used so it has been removed.
---
 .../clones/connectivity_and_data.py           | 130 ++++++++----------
 .../pki/server/healthcheck/clones/plugin.py   |  75 ++++------
 base/server/python/pki/server/__init__.py     |   8 +-
 3 files changed, 91 insertions(+), 122 deletions(-)

diff --git a/base/server/healthcheck/pki/server/healthcheck/clones/connectivity_and_data.py b/base/server/healthcheck/pki/server/healthcheck/clones/connectivity_and_data.py
index ca5d6dae48..d9bb480f7f 100644
--- a/base/server/healthcheck/pki/server/healthcheck/clones/connectivity_and_data.py
+++ b/base/server/healthcheck/pki/server/healthcheck/clones/connectivity_and_data.py
@@ -46,93 +46,83 @@ class ClonesConnectivyAndDataCheck(ClonesPlugin):
 
     def check_kra_clones(self):
         for host in self.clone_kras:
-            cur_clone_msg = ' Host: ' + host.Hostname + ' Port: ' + host.SecurePort
-            # Reach out and get some keys or requests , to serve as a data and connectivity check
+
+            url = 'https://' + host.Hostname + ':' + host.SecurePort
+
             try:
-                client_nick = self.security_domain.config.get('ca.connector.KRA.nickName')
-
-                output = self.contact_subsystem_using_pki(
-                    host.SecurePort, host.Hostname, client_nick,
-                    self.passwd, self.db_dir, 'kra-key-show', ['0x01'])
-
-                # check to see if we either got a key or a key not found exception
-                # of which either will imply a successful connection
-                if output is not None:
-                    key_found = output.find('Key ID:')
-                    key_not_found = output.find('KeyNotFoundException:')
-                    if key_found >= 0:
-                        logger.info('Key material found from kra clone.')
-
-                    if key_not_found >= 0:
-                        logger.info('key not found, possibly empty kra')
-
-                    if key_not_found == -1 and key_found == -1:
-                        logger.info('Failure to get key material from kra')
-                        raise BaseException('KRA clone problem detected ' + cur_clone_msg)
-                else:
-                    raise BaseException('No data obtained from KRA clone.' + cur_clone_msg)
+                status = self.get_status(
+                    host.Hostname,
+                    host.SecurePort,
+                    '/kra/admin/kra/getStatus')
 
-            except BaseException as e:
-                logger.error("Internal error testing KRA clone. %s", e)
-                raise BaseException('Internal error testing KRA clone.' + cur_clone_msg)
+                logger.info('KRA at %s is %s', url, status)
 
-        return
+                if status != 'running':
+                    raise Exception('KRA at %s is %s' % (url, status))
+
+            except Exception as e:
+                logger.error('Unable to reach KRA at %s: %s', url, e)
+                raise Exception('Unable to reach KRA at %s: %s' % (url, e))
 
     def check_ocsp_clones(self):
         for host in self.clone_ocsps:
-            cur_clone_msg = ' Host: ' + host.Hostname + ' Port: ' + host.SecurePort
-            # Reach out to the ocsp clones
+
+            url = 'https://' + host.Hostname + ':' + host.SecurePort
+
             try:
-                output = self.contact_subsystem_using_sslget(
-                    host.SecurePort, host.Hostname, None,
-                    self.passwd, self.db_dir, None, '/ocsp/admin/ocsp/getStatus')
-
-                good_status = output.find('<State>1</State>')
-                if good_status == -1:
-                    raise BaseException('OCSP clone problem detected.' + cur_clone_msg)
-                logger.info('good_status %s ', good_status)
-            except BaseException as e:
-                logger.error("Internal error testing OCSP clone.  %s", e)
-                raise BaseException('Internal error testing OCSP clone.' + cur_clone_msg)
+                status = self.get_status(
+                    host.Hostname,
+                    host.SecurePort,
+                    '/ocsp/admin/ocsp/getStatus')
 
-        return
+                logger.info('OCSP at %s is %s', url, status)
+
+                if status != 'running':
+                    raise Exception('OCSP at %s is %s' % (url, status))
+
+            except Exception as e:
+                logger.error('Unable to reach OCSP at %s: %s', url, e)
+                raise Exception('Unable to reach OCSP at %s: %s' % (url, e))
 
     def check_tks_clones(self):
         for host in self.clone_tkss:
-            cur_clone_msg = ' Host: ' + host.Hostname + ' Port: ' + host.SecurePort
-            # Reach out to the tks clones
+
+            url = 'https://' + host.Hostname + ':' + host.SecurePort
+
             try:
-                output = self.contact_subsystem_using_sslget(
-                    host.SecurePort, host.Hostname, None,
-                    self.passwd, self.db_dir, None, '/tks/admin/tks/getStatus')
-
-                good_status = output.find('<State>1</State>')
-                if good_status == -1:
-                    raise BaseException('TKS clone problem detected.' + cur_clone_msg)
-                logger.info('good_status %s ', good_status)
-            except BaseException as e:
-                logger.error("Internal error testing TKS clone. %s", e)
-                raise BaseException('Internal error testing TKS clone.' + cur_clone_msg)
+                status = self.get_status(
+                    host.Hostname,
+                    host.SecurePort,
+                    '/tks/admin/tks/getStatus')
 
-        return
+                logger.info('TKS at %s is %s', url, status)
+
+                if status != 'running':
+                    raise Exception('TKS at %s is %s' % (url, status))
+
+            except Exception as e:
+                logger.error('Unable to reach TKS at %s: %s', url, e)
+                raise Exception('Unable to reach TKS at %s: %s' % (url, e))
 
     def check_tps_clones(self):
         for host in self.clone_tpss:
-            cur_clone_msg = ' Host: ' + host.Hostname + ' Port: ' + host.SecurePort
-            # Reach out to the tps clones
+
+            url = 'https://' + host.Hostname + ':' + host.SecurePort
+
             try:
-                output = self.contact_subsystem_using_sslget(
-                    host.SecurePort, host.Hostname, None,
-                    self.passwd, self.db_dir, None, '/tps/admin/tps/getStatus')
-
-                good_status = output.find('<State>1</State>')
-                if good_status == -1:
-                    raise BaseException('TPS clone problem detected.' + cur_clone_msg)
-                logger.info('good_status  %s ', good_status)
-            except BaseException as e:
-                logger.error("Internal error testing TPS clone. %s", e)
-                raise BaseException('Internal error testing TPS clone.' + cur_clone_msg)
-        return
+                status = self.get_status(
+                    host.Hostname,
+                    host.SecurePort,
+                    '/tps/admin/tps/getStatus')
+
+                logger.info('TPS at %s is %s', url, status)
+
+                if status != 'running':
+                    raise Exception('TPS at %s is %s' % (url, status))
+
+            except Exception as e:
+                logger.error('Unable to reach TPS at %s: %s', url, e)
+                raise Exception('Unable to reach TPS at %s: %s' % (url, e))
 
     @duration
     def check(self):
diff --git a/base/server/healthcheck/pki/server/healthcheck/clones/plugin.py b/base/server/healthcheck/pki/server/healthcheck/clones/plugin.py
index 2472f35b5b..824c36a1a9 100644
--- a/base/server/healthcheck/pki/server/healthcheck/clones/plugin.py
+++ b/base/server/healthcheck/pki/server/healthcheck/clones/plugin.py
@@ -6,6 +6,10 @@
 # SPDX-License-Identifier: GPL-2.0-or-later
 #
 
+import json
+import logging
+import xml.etree.ElementTree as ET
+
 from ipahealthcheck.core.plugin import Plugin, Registry
 from pki.server.instance import PKIInstance
 from pki.client import PKIConnection
@@ -13,9 +17,6 @@ from pki.system import SecurityDomainClient
 
 from pki.server.healthcheck.core.main import merge_dogtag_config
 
-import logging
-import subprocess
-
 logger = logging.getLogger(__name__)
 
 # Temporary workaround to skip VERBOSE data. Fix already pushed to upstream
@@ -46,60 +47,36 @@ class ClonesPlugin(Plugin):
 
         self.instance = PKIInstance(self.config.instance_name)
 
-    def contact_subsystem_using_pki(
-            self, subport, subhost, subsystemnick,
-            token_pwd, db_path, cmd, exts=None):
-        command = ["/usr/bin/pki",
-                   "-p", str(subport),
-                   "-h", subhost,
-                   "-n", subsystemnick,
-                   "-P", "https",
-                   "-d", db_path,
-                   "-c", token_pwd,
-                   cmd]
-
-        if exts is not None:
-            command.extend(exts)
-
-        output = None
-        try:
-            output = subprocess.check_output(command, stderr=subprocess.STDOUT)
-        except subprocess.CalledProcessError as e:
-            output = e.output.decode('utf-8')
-            return output
+    def get_status(self, host, port, path):
 
-        output = output.decode('utf-8')
+        self.instance.export_ca_cert()
 
-        return output
+        connection = PKIConnection(
+            protocol='https',
+            hostname=host,
+            port=port,
+            cert_paths=self.instance.ca_cert)
 
-    def contact_subsystem_using_sslget(
-            self, port, host, subsystemnick,
-            token_pwd, db_path, params, url):
+        response = connection.get(path)
 
-        command = ["/usr/bin/sslget"]
+        content_type = response.headers['Content-Type']
+        content = response.text
+        logger.info('Content:\n%s', content)
 
-        if subsystemnick is not None:
-            command.extend(["-n", subsystemnick])
+        # https://github.com/dogtagpki/pki/wiki/GetStatus-Service
+        if content_type == 'application/json':
+            json_response = json.loads(content)
+            status = json_response['Response']['Status']
 
-        command.extend(["-p", token_pwd, "-d", db_path])
-
-        if params is not None:
-            command.extend(["-e", params])
-
-        command.extend([
-            "-r", url, host + ":" + port])
-
-        logger.info(' command : %s ', command)
-        output = None
-        try:
-            output = subprocess.check_output(command, stderr=subprocess.STDOUT)
-        except subprocess.CalledProcessError as e:
-            output = e.output.decode('utf-8')
-            return output
+        elif content_type == 'application/xml':
+            root = ET.fromstring(content)
+            status = root.findtext('Status')
 
-        output = output.decode('utf-8')
+        else:
+            raise Exception('Unsupported content-type: %s' % content_type)
 
-        return output
+        logger.info('Status: %s', status)
+        return status
 
     def get_security_domain_data(self, host, port):
         domain_data = None
diff --git a/base/server/python/pki/server/__init__.py b/base/server/python/pki/server/__init__.py
index 4fbb74684b..0515bbb197 100644
--- a/base/server/python/pki/server/__init__.py
+++ b/base/server/python/pki/server/__init__.py
@@ -241,6 +241,10 @@ class PKIServer(object):
     def jss_conf(self):
         return os.path.join(self.conf_dir, 'jss.conf')
 
+    @property
+    def ca_cert(self):
+        return os.path.join(self.nssdb_dir, 'ca.crt')
+
     def is_valid(self):
         return self.exists()
 
@@ -259,8 +263,6 @@ class PKIServer(object):
 
     def export_ca_cert(self):
 
-        ca_path = os.path.join(self.nssdb_dir, 'ca.crt')
-
         token = pki.nssdb.INTERNAL_TOKEN_NAME
         nickname = self.get_sslserver_cert_nickname()
 
@@ -272,7 +274,7 @@ class PKIServer(object):
         nssdb = self.open_nssdb(token=token)
 
         try:
-            nssdb.extract_ca_cert(ca_path, nickname)
+            nssdb.extract_ca_cert(self.ca_cert, nickname)
         finally:
             nssdb.close()
 
-- 
2.33.1