Script 'mail_helper' called by obssrc
Hello community,
here is the log from the commit of package crmsh for openSUSE:Factory checked
in at 2025-04-20 09:35:10
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/crmsh (Old)
and /work/SRC/openSUSE:Factory/.crmsh.new.30101 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "crmsh"
Sun Apr 20 09:35:10 2025 rev:364 rq:1270864 version:5.0.0+20250418.bb6c26c2
Thu Apr 17 16:10:12 2025 rev:363 rq:1270229 version:5.0.0+20250417.367bbfad
Changes:
--------
--- /work/SRC/openSUSE:Factory/crmsh/crmsh.changes 2025-04-03
16:51:53.631970717 +0200
+++ /work/SRC/openSUSE:Factory/.crmsh.new.30101/crmsh.changes 2025-04-20
19:51:02.262549551 +0200
@@ -1,0 +2,24 @@
+Fri Apr 18 14:42:07 UTC 2025 - xli...@suse.com
+
+- Update to version 5.0.0+20250418.bb6c26c2:
+ * Dev: unittests: Adjust unit test for previous commit
+ * Dev: ui_sbd: Improve log info when adjust sbd related timeout
+ * Dev: doc: Update crm.8.adoc for using -F/--force option for leveraging
maintenance mode
+ * Dev: sbd: Leverage maintenance mode when need to restart cluster
(jsc#PED-11931)
+
+-------------------------------------------------------------------
+Thu Apr 17 08:27:07 UTC 2025 - xli...@suse.com
+
+- Update to version 5.0.0+20250417.367bbfad:
+ * Dev: unittests: Adjust unit test for previous commit
+ * Dev: behave: Adjust functional test for previous commit
+ * Dev: ui_sbd: Show fence_sbd parameter in 'sbd configure show'
+ * Dev: doc: Update doc/crm.8.adoc to add crashdump option
+ * Dev: ui_sbd: Add 'crashdump' option for 'sbd purge' command
+ * Dev: ui_sbd: Refactor the condition for configuring crashdump
+ * Dev: ui_sbd: Compare crashdump watchdog timeout value if configured
+ * Dev: ui_sbd: Add warning to emphasize that kdump service is required
+ * Dev: ui_sbd: Update the re pattern to match the possible arguments
+ * Dev: ui_sbd: Configure crashdump watchdog timeout (jsc#PED-11931)
+
+-------------------------------------------------------------------
Old:
----
crmsh-5.0.0+20250403.1442a04a.tar.bz2
New:
----
crmsh-5.0.0+20250418.bb6c26c2.tar.bz2
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Other differences:
------------------
++++++ crmsh.spec ++++++
--- /var/tmp/diff_new_pack.3fAZbx/_old 2025-04-20 19:51:04.222631123 +0200
+++ /var/tmp/diff_new_pack.3fAZbx/_new 2025-04-20 19:51:04.238631789 +0200
@@ -36,7 +36,7 @@
Summary: High Availability cluster command-line interface
License: GPL-2.0-or-later
Group: %{pkg_group}
-Version: 5.0.0+20250403.1442a04a
+Version: 5.0.0+20250418.bb6c26c2
Release: 0
URL: http://crmsh.github.io
Source0: %{name}-%{version}.tar.bz2
++++++ _servicedata ++++++
--- /var/tmp/diff_new_pack.3fAZbx/_old 2025-04-20 19:51:04.602646938 +0200
+++ /var/tmp/diff_new_pack.3fAZbx/_new 2025-04-20 19:51:04.650648936 +0200
@@ -9,7 +9,7 @@
</service>
<service name="tar_scm">
<param name="url">https://github.com/ClusterLabs/crmsh.git</param>
- <param
name="changesrevision">1442a04aba88f012c0995982e64ed9589f140a45</param>
+ <param
name="changesrevision">bfb056b1c60929d7b97d9ee1c37d0e3714c60d46</param>
</service>
</servicedata>
(No newline at EOF)
++++++ crmsh-5.0.0+20250403.1442a04a.tar.bz2 ->
crmsh-5.0.0+20250418.bb6c26c2.tar.bz2 ++++++
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/crmsh/bootstrap.py
new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/bootstrap.py
--- old/crmsh-5.0.0+20250403.1442a04a/crmsh/bootstrap.py 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/bootstrap.py 2025-04-18
16:21:27.000000000 +0200
@@ -2725,12 +2725,12 @@
for res in cib_factory.fence_id_list_without_pcmk_delay():
cmd = "crm resource param {} set pcmk_delay_max {}s".format(res,
PCMK_DELAY_MAX)
shell.get_stdout_or_raise_error(cmd)
- logger.debug("Add parameter 'pcmk_delay_max={}s' for resource
'{}'".format(PCMK_DELAY_MAX, res))
+ logger.info("Add parameter 'pcmk_delay_max={}s' for resource
'{}'".format(PCMK_DELAY_MAX, res))
else:
for res in cib_factory.fence_id_list_with_pcmk_delay():
cmd = "crm resource param {} delete pcmk_delay_max".format(res)
shell.get_stdout_or_raise_error(cmd)
- logger.debug("Delete parameter 'pcmk_delay_max' for resource
'{}'".format(res))
+ logger.info("Delete parameter 'pcmk_delay_max' for resource
'{}'".format(res))
def adjust_stonith_timeout():
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/crmsh/cibquery.py
new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/cibquery.py
--- old/crmsh-5.0.0+20250403.1442a04a/crmsh/cibquery.py 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/cibquery.py 2025-04-18
16:21:27.000000000 +0200
@@ -33,6 +33,26 @@
f'/instance_attributes/nvpair[@name="fstype" and @value="{fstype}"]'
))
+
+def get_primitives_with_ra(cib: lxml.etree.Element, ra: ResourceAgent) ->
list[str]:
+ """
+ Given cib and ResourceAgent instance, return id list of primitives that
matched
+ consider provider as optional
+ """
+ provider_condition = f' and @provider="{ra.m_provider}"' if ra.m_provider
else ""
+ return cib.xpath(
+
f'/cib/configuration/resources//primitive[@class="{ra.m_class}"{provider_condition}
and @type="{ra.m_type}"]/@id'
+ )
+
+
+def get_parameter_value(cib: lxml.etree.Element, res_id: str, param_name: str)
-> typing.Optional[str]:
+ result = cib.xpath(
+ f'/cib/configuration/resources//primitive[@id="{res_id}"]'
+ f'/instance_attributes/nvpair[@name="{param_name}"]/@value'
+ )
+ return result[0] if result else None
+
+
def get_cluster_nodes(cib: lxml.etree.Element) -> list[ClusterNode]:
"""Return a list of cluster nodes, excluding pacemaker-remote nodes"""
result = list()
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/crmsh/sbd.py
new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/sbd.py
--- old/crmsh-5.0.0+20250403.1442a04a/crmsh/sbd.py 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/sbd.py 2025-04-18
16:21:27.000000000 +0200
@@ -81,6 +81,14 @@
return utils.parse_sysconfig(SBDManager.SYSCONFIG_SBD).get(key)
@staticmethod
+ def get_crashdump_watchdog_timeout() -> typing.Optional[int]:
+ res = SBDUtils.get_sbd_value_from_config("SBD_OPTS")
+ if not res:
+ return None
+ matched = re.search(r"-C\s+(\d+)", res)
+ return int(matched.group(1)) if matched else None
+
+ @staticmethod
def get_sbd_device_from_config():
'''
Get sbd device list from config
@@ -546,11 +554,13 @@
service_manager.enable_service(constants.SBD_SERVICE, node)
@staticmethod
- def restart_cluster_if_possible():
+ def restart_cluster_if_possible(with_maintenance_mode=False):
if not ServiceManager().service_is_active(constants.PCMK_SERVICE):
return
- if xmlutil.CrmMonXmlParser().is_any_resource_running():
+ if xmlutil.CrmMonXmlParser().is_any_resource_running() and not
with_maintenance_mode:
logger.warning("Resource is running, need to restart cluster
service manually on each node")
+ logger.warning("Or, run with `crm -F` or `--force` option, the
`sbd` subcommand will leverage maintenance mode for any changes that require
restarting sbd.service")
+ logger.warning("Understand risks that running RA has no cluster
protection while the cluster is in maintenance mode and restarting")
else:
bootstrap.restart_cluster()
@@ -559,7 +569,8 @@
Configure fence_sbd resource and related properties
'''
if self.diskless_sbd:
- utils.set_property("stonith-watchdog-timeout",
SBDTimeout.STONITH_WATCHDOG_TIMEOUT_DEFAULT)
+ swt_value = self.timeout_dict.get("stonith-watchdog",
SBDTimeout.STONITH_WATCHDOG_TIMEOUT_DEFAULT)
+ utils.set_property("stonith-watchdog-timeout", swt_value)
else:
if utils.get_property("stonith-watchdog-timeout",
get_default=False):
utils.delete_property("stonith-watchdog-timeout")
@@ -684,14 +695,15 @@
return
self._load_attributes_from_bootstrap()
- self.initialize_sbd()
- self.update_configuration()
- SBDManager.enable_sbd_service()
-
- if self.cluster_is_running:
- SBDManager.restart_cluster_if_possible()
- self.configure_sbd()
- bootstrap.adjust_properties()
+ with utils.leverage_maintenance_mode() as enabled:
+ self.initialize_sbd()
+ self.update_configuration()
+ SBDManager.enable_sbd_service()
+
+ if self.cluster_is_running:
+
SBDManager.restart_cluster_if_possible(with_maintenance_mode=enabled)
+ self.configure_sbd()
+ bootstrap.adjust_properties()
def join_sbd(self, remote_user, peer_host):
'''
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/crmsh/ui_sbd.py
new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/ui_sbd.py
--- old/crmsh-5.0.0+20250403.1442a04a/crmsh/ui_sbd.py 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/ui_sbd.py 2025-04-18
16:21:27.000000000 +0200
@@ -12,6 +12,7 @@
from crmsh import sh
from crmsh import xmlutil
from crmsh import constants
+from crmsh import cibquery
from crmsh.service_manager import ServiceManager
@@ -88,30 +89,37 @@
- sbd purge
'''
name = "sbd"
- TIMEOUT_TYPES = ("watchdog", "allocate", "loop", "msgwait")
- DISKLESS_TIMEOUT_TYPES = ("watchdog",)
+ TIMEOUT_TYPES = ("watchdog", "allocate", "loop", "msgwait",
"crashdump-watchdog")
+ DISKLESS_TIMEOUT_TYPES = ("watchdog", "crashdump-watchdog")
SHOW_TYPES = ("disk_metadata", "sysconfig", "property")
DISKLESS_SHOW_TYPES = ("sysconfig", "property")
PCMK_ATTRS = (
"have-watchdog",
"stonith-timeout",
- "stonith-enabled",
- "priority-fencing-delay",
- "pcmk_delay_max"
+ "stonith-enabled"
)
PCMK_ATTRS_DISKLESS = ('stonith-watchdog-timeout',)
PARSE_RE = re.compile(
- # Match keys with non-empty values, capturing possible suffix
- r'(\w+)(?:-(\w+))?=("[^"]+"|[\w/\d;]+)'
+ # To extract key, suffix and value from these possible arguments:
+ # watchdog-timeout=30
+ # crashdump-watchdog-timeout=120
+ # watchdog-device=/dev/watchdog
+ r'([\w-]+)-([\w]+)=([\w/]+)'
)
+ # re pattern to match "-C <number>" or "-C <number> -Z"
+ SBD_OPTS_RE = r'-C\s+\d+(\s+-Z)?'
class SyntaxError(Exception):
pass
+ class MissingRequiredException(Exception):
+ pass
+
def __init__(self):
self.device_list_from_config: list[str] = None
self.device_meta_dict_runtime: dict[str, int] = None
self.watchdog_timeout_from_config: int = None
+ self.crashdump_watchdog_timeout_from_config: int = None
self.watchdog_device_from_config: str = None
self.service_manager: ServiceManager = None
self.cluster_shell: sh.cluster_shell = None
@@ -130,6 +138,7 @@
except Exception:
self.watchdog_timeout_from_config = None
self.watchdog_device_from_config =
watchdog.Watchdog.get_watchdog_device_from_sbd_config()
+ self.crashdump_watchdog_timeout_from_config =
sbd.SBDUtils.get_crashdump_watchdog_timeout()
self.service_manager = ServiceManager()
self.cluster_shell = sh.cluster_shell()
@@ -217,6 +226,13 @@
for match in matches:
print(f"{match[0]}={match[1]}")
+ cmd = "crm configure show related:fence_sbd"
+ out = self.cluster_shell.get_stdout_or_raise_error(cmd)
+ if out:
+ print()
+ logger.info('%s', cmd)
+ print(out)
+
print()
logger.info('%s', sbd.SBDTimeout.SHOW_SBD_START_TIMEOUT_CMD)
systemd_start_timeout = sbd.SBDTimeout.get_sbd_systemd_start_timeout()
@@ -280,13 +296,100 @@
return timeout_dict
if watchdog_timeout and not msgwait_timeout:
timeout_dict["msgwait"] = 2*watchdog_timeout
- logger.info("No msgwait timeout specified, use 2*watchdog timeout:
%s", 2*watchdog_timeout)
+ logger.info("No 'msgwait-timeout=' specified in the command, use
2*watchdog timeout: %s", 2*watchdog_timeout)
return timeout_dict
if msgwait_timeout and not watchdog_timeout:
watchdog_timeout = msgwait_timeout//2
timeout_dict["watchdog"] = watchdog_timeout
- logger.info("No watchdog timeout specified, use msgwait timeout/2:
%s", watchdog_timeout)
+ logger.info("No 'watchdog-timeout=' specified in the command, use
msgwait timeout/2: %s", watchdog_timeout)
return timeout_dict
+ return timeout_dict
+
+ def _set_crashdump_option(self, delete=False):
+ '''
+ Set crashdump option for fence_sbd resource
+ '''
+ cib =
xmlutil.text2elem(self.cluster_shell.get_stdout_or_raise_error('crm configure
show xml'))
+ ra = cibquery.ResourceAgent("stonith", "", "fence_sbd")
+ res_id_list = cibquery.get_primitives_with_ra(cib, ra)
+ if not res_id_list:
+ if delete:
+ return
+ logger.error("No fence_sbd resource found")
+ raise self.MissingRequiredException
+
+ crashdump_value = cibquery.get_parameter_value(cib, res_id_list[0],
"crashdump")
+ cmd = ""
+ if utils.is_boolean_false(crashdump_value):
+ if delete:
+ return
+ cmd = f"crm resource param {res_id_list[0]} set crashdump 1"
+ logger.info("Set crashdump option for fence_sbd resource")
+ elif delete:
+ cmd = f"crm resource param {res_id_list[0]} delete crashdump"
+ logger.info("Delete crashdump option for fence_sbd resource")
+ if cmd:
+ self.cluster_shell.get_stdout_or_raise_error(cmd)
+
+ def _set_crashdump_in_sysconfig(self, crashdump_watchdog_timeout=None,
restore=False, diskless=False) -> dict:
+ update_dict = {}
+ sbd_timeout_action_for_crashdump = "flush,crashdump"
+ comment_action_line = f"sed -i
'/^SBD_TIMEOUT_ACTION/s/^/#__sbd_crashdump_backup__ /'
{sbd.SBDManager.SYSCONFIG_SBD}"
+ add_action_line = f"sed -i '/^#__sbd_crashdump_backup__/a
SBD_TIMEOUT_ACTION={sbd_timeout_action_for_crashdump}'
{sbd.SBDManager.SYSCONFIG_SBD}"
+ comment_out_action_line = f"sed -i 's/^#__sbd_crashdump_backup__
SBD_TIMEOUT_ACTION/SBD_TIMEOUT_ACTION/' {sbd.SBDManager.SYSCONFIG_SBD}"
+ delete_action_line = f"sed -i '/^SBD_TIMEOUT_ACTION/d'
{sbd.SBDManager.SYSCONFIG_SBD}"
+
+ sbd_timeout_action_configured =
sbd.SBDUtils.get_sbd_value_from_config("SBD_TIMEOUT_ACTION")
+ if restore:
+ if sbd_timeout_action_configured and sbd_timeout_action_configured
== sbd_timeout_action_for_crashdump:
+ cmd_delete_and_comment_out = f"{delete_action_line} &&
{comment_out_action_line}"
+ logger.info("Delete SBD_TIMEOUT_ACTION: %s and restore
original value", sbd_timeout_action_for_crashdump)
+
self.cluster_shell.get_stdout_or_raise_error(cmd_delete_and_comment_out)
+
+ sbd_opts = sbd.SBDUtils.get_sbd_value_from_config("SBD_OPTS")
+ if sbd_opts and re.search(self.SBD_OPTS_RE, sbd_opts):
+ sbd_opts = re.sub(self.SBD_OPTS_RE, '', sbd_opts)
+ update_dict["SBD_OPTS"] = ' '.join(sbd_opts.split())
+
+ elif crashdump_watchdog_timeout:
+ if not sbd_timeout_action_configured:
+ update_dict["SBD_TIMEOUT_ACTION"] =
sbd_timeout_action_for_crashdump
+ elif sbd_timeout_action_configured !=
sbd_timeout_action_for_crashdump:
+ cmd_comment_and_add = f"{comment_action_line} &&
{add_action_line}"
+
self.cluster_shell.get_stdout_or_raise_error(cmd_comment_and_add)
+ logger.info("Update SBD_TIMEOUT_ACTION in %s: %s",
sbd.SBDManager.SYSCONFIG_SBD, sbd_timeout_action_for_crashdump)
+
+ value_for_diskless = " -Z" if diskless else ""
+ value_for_sbd_opts = f"-C
{crashdump_watchdog_timeout}{value_for_diskless}"
+ sbd_opts = sbd.SBDUtils.get_sbd_value_from_config("SBD_OPTS")
+ sbd_opts = re.sub(self.SBD_OPTS_RE, '', sbd_opts)
+ update_dict["SBD_OPTS"] = f"{' '.join(sbd_opts.split())}
{value_for_sbd_opts}" if sbd_opts else value_for_sbd_opts
+
+ return update_dict
+
+ def _check_kdump_service(self):
+ no_kdump = False
+ for node in self.cluster_nodes:
+ if not self.service_manager.service_is_active("kdump.service",
node):
+ logger.warning("Kdump service is not active on %s", node)
+ no_kdump = True
+ if no_kdump:
+ logger.warning("Kdump service is required for crashdump")
+
+ def _should_configure_crashdump(
+ self,
+ crashdump_watchdog_timeout,
+ watchdog_timeout,
+ diskless=False
+ ) -> bool:
+ if not crashdump_watchdog_timeout and not
self.crashdump_watchdog_timeout_from_config:
+ return False
+ ct_updated = crashdump_watchdog_timeout and \
+ crashdump_watchdog_timeout !=
self.crashdump_watchdog_timeout_from_config
+ watchdog_timeout_configured = self.watchdog_timeout_from_config if
diskless \
+ else self.device_meta_dict_runtime.get("watchdog")
+ wt_updated = watchdog_timeout and watchdog_timeout !=
watchdog_timeout_configured
+ return ct_updated or wt_updated
def _configure_diskbase(self, parameter_dict: dict):
'''
@@ -296,18 +399,28 @@
watchdog_device = parameter_dict.get("watchdog-device")
if watchdog_device != self.watchdog_device_from_config:
update_dict["SBD_WATCHDOG_DEV"] = watchdog_device
- timeout_dict = {k: v for k, v in parameter_dict.items() if k in
self.TIMEOUT_TYPES}
- is_subdict_timeout = utils.is_subdict(timeout_dict,
self.device_meta_dict_runtime)
- if is_subdict_timeout and not update_dict:
+ timeout_dict = {
+ k: v for k, v in parameter_dict.items()
+ if k in self.TIMEOUT_TYPES and k != "crashdump-watchdog"
+ }
+ timeout_dict = self._adjust_timeout_dict(timeout_dict)
+ # merge runtime timeout dict into parameter timeout dict without
overwriting
+ timeout_dict = {**self.device_meta_dict_runtime, **timeout_dict}
+
+ crashdump_watchdog_timeout = parameter_dict.get("crashdump-watchdog",
self.crashdump_watchdog_timeout_from_config)
+ if self._should_configure_crashdump(crashdump_watchdog_timeout,
timeout_dict.get("watchdog")):
+ self._check_kdump_service()
+ self._set_crashdump_option()
+ timeout_dict["msgwait"] = 2*timeout_dict["watchdog"] +
crashdump_watchdog_timeout
+ logger.info("Set msgwait-timeout to 2*watchdog-timeout +
crashdump-watchdog-timeout: %s", timeout_dict["msgwait"])
+ result_dict =
self._set_crashdump_in_sysconfig(crashdump_watchdog_timeout)
+ update_dict = {**update_dict, **result_dict}
+
+ if timeout_dict == self.device_meta_dict_runtime and not update_dict:
logger.info("No change in SBD configuration")
return
- if not is_subdict_timeout:
- timeout_dict = self._adjust_timeout_dict(timeout_dict)
- # merge runtime timeout dict into parameter timeout dict without
overwriting
- timeout_dict = {**self.device_meta_dict_runtime, **timeout_dict}
-
sbd_manager = sbd.SBDManager(
device_list_to_init=self.device_list_from_config,
timeout_dict=timeout_dict,
@@ -320,17 +433,30 @@
Configure diskless SBD based on input parameters and runtime config
'''
update_dict = {}
+ timeout_dict = {}
+
watchdog_timeout = parameter_dict.get("watchdog")
if watchdog_timeout and watchdog_timeout !=
self.watchdog_timeout_from_config:
update_dict["SBD_WATCHDOG_TIMEOUT"] = str(watchdog_timeout)
watchdog_device = parameter_dict.get("watchdog-device")
if watchdog_device != self.watchdog_device_from_config:
update_dict["SBD_WATCHDOG_DEV"] = watchdog_device
+
+ crashdump_watchdog_timeout = parameter_dict.get("crashdump-watchdog",
self.crashdump_watchdog_timeout_from_config)
+ if self._should_configure_crashdump(crashdump_watchdog_timeout,
watchdog_timeout, diskless=True):
+ self._check_kdump_service()
+ result_dict =
self._set_crashdump_in_sysconfig(crashdump_watchdog_timeout, diskless=True)
+ update_dict = {**update_dict, **result_dict}
+ sbd_watchdog_timeout = watchdog_timeout or
self.watchdog_timeout_from_config
+ stonith_watchdog_timeout = sbd_watchdog_timeout +
crashdump_watchdog_timeout
+ logger.info("Set stonith-watchdog-timeout to SBD_WATCHDOG_TIMEOUT
+ crashdump-watchdog-timeout: %s", stonith_watchdog_timeout)
+ timeout_dict["stonith-watchdog"] = stonith_watchdog_timeout
if not update_dict:
logger.info("No change in SBD configuration")
return
sbd_manager = sbd.SBDManager(
+ timeout_dict=timeout_dict,
update_dict=update_dict,
diskless_sbd=True
)
@@ -379,6 +505,7 @@
'''
Implement sbd device command
'''
+ self._load_attributes()
if not self.service_is_active(constants.PCMK_SERVICE):
return False
if not sbd.SBDUtils.is_using_disk_based_sbd():
@@ -417,15 +544,16 @@
Implement sbd configure command
'''
try:
- for service in (constants.PCMK_SERVICE, constants.SBD_SERVICE):
- if not self.service_is_active(service):
- return False
+ self._load_attributes()
if not args:
raise self.SyntaxError("No argument")
-
if args[0] == "show":
self._configure_show(args)
return True
+ for service in (constants.PCMK_SERVICE, constants.SBD_SERVICE):
+ if not self.service_is_active(service):
+ return False
+
parameter_dict = self._parse_args(args)
if sbd.SBDUtils.is_using_disk_based_sbd():
self._configure_diskbase(parameter_dict)
@@ -439,13 +567,26 @@
if usage:
print(usage)
return False
+ except self.MissingRequiredException:
+ return False
- def do_purge(self, context) -> bool:
+ @command.completers(completers.choice(['crashdump']))
+ def do_purge(self, context, *args) -> bool:
'''
Implement sbd purge command
'''
+ self._load_attributes()
if not self.service_is_active(constants.SBD_SERVICE):
return False
+
+ if args and args[0] == "crashdump":
+ self._set_crashdump_option(delete=True)
+ update_dict = self._set_crashdump_in_sysconfig(restore=True)
+ if update_dict:
+ sbd.SBDManager.update_sbd_configuration(update_dict)
+ sbd.SBDManager.restart_cluster_if_possible()
+ return True
+
sbd.purge_sbd_from_cluster()
sbd.SBDManager.restart_cluster_if_possible()
return True
@@ -544,6 +685,7 @@
'''
Implement sbd status command
'''
+ self._load_attributes()
self._print_sbd_type()
self._print_sbd_status()
self._print_sbd_cgroup_status()
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/crmsh/utils.py
new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/utils.py
--- old/crmsh-5.0.0+20250403.1442a04a/crmsh/utils.py 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/crmsh/utils.py 2025-04-18
16:21:27.000000000 +0200
@@ -976,6 +976,20 @@
return False
+def is_dc_idle():
+ dc = get_dc()
+ if not dc:
+ return False
+ cmd = f"crmadmin -S {dc}"
+ rc, out, err = ShellUtils().get_stdout_stderr(cmd)
+ if rc != 0 and err:
+ logger.error("Failed to get DC status: %s", err)
+ return False
+ if not out:
+ return False
+ return "ok" in out and "S_IDLE" in out
+
+
def get_dc(peer=None):
cmd = "crmadmin -D -t 1"
_, out, _ = sh.cluster_shell().get_rc_stdout_stderr_without_input(peer,
cmd)
@@ -2788,6 +2802,25 @@
return False
+@contextmanager
+def leverage_maintenance_mode() -> bool:
+ if not config.core.force:
+ yield False
+ return
+
+ if is_dc_idle():
+ try:
+ logger.info("Set cluster to maintenance mode")
+ set_property("maintenance-mode", "true")
+ yield True
+ finally:
+ logger.info("Set cluster from maintenance mode to normal")
+ delete_property("maintenance-mode")
+ else:
+ logger.warning("Pacemaker state transition is in progress. Skip
restarting cluster in maintenance mode.")
+ yield False
+
+
def check_no_quorum_policy_with_dlm():
"""
Give warning when no-quorum-policy not freeze while configured DLM
@@ -3193,11 +3226,4 @@
"""
ansi_escape_pattern = re.compile(r'\x1B\[[0-?]*[ -/]*[@-~]')
return ansi_escape_pattern.sub('', text)
-
-
-def is_subdict(sub_dict, main_dict):
- """
- Check if sub_dict is a sub-dictionary of main_dict
- """
- return all(main_dict.get(k) == v for k, v in sub_dict.items())
# vim:ts=4:sw=4:et:
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore' old/crmsh-5.0.0+20250403.1442a04a/doc/crm.8.adoc
new/crmsh-5.0.0+20250418.bb6c26c2/doc/crm.8.adoc
--- old/crmsh-5.0.0+20250403.1442a04a/doc/crm.8.adoc 2025-04-03
10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/doc/crm.8.adoc 2025-04-18
16:21:27.000000000 +0200
@@ -2126,6 +2126,11 @@
as well as the on-disk metadata for the disk-based scenario.
Currently, SBD management requires a running cluster.
+When run with `crm -F` or `--force` option, the `sbd` subcommand will leverage
maintenance mode
+for any changes that require restarting sbd.service.
+WARNING: Understand risks that running RA has no cluster protection while the
cluster is
+in maintenance mode and restarting
+
[[cmdhelp.sbd.configure,configure SBD]]
==== `configure`
@@ -2136,6 +2141,7 @@
- Show contents of /etc/sysconfig/sbd
- Show SBD related cluster properties
- Update the SBD related configuration parameters
+- NOTE: sbd crashdump is used for debugging. Understand the risks and run `crm
sbd purge crashdump` afterward
For more details on SBD and related parameters, please see man sbd(8).
@@ -2143,11 +2149,11 @@
...............
# For disk-based SBD
crm sbd configure show [disk_metadata|sysconfig|property]
-crm sbd configure [watchdog-timeout=<integer>] [allocate-timeout=<integer>]
[loop-timeout=<integer>] [msgwait-timeout=<integer>] [watchdog-device=<device>]
+crm sbd configure [watchdog-timeout=<integer>] [allocate-timeout=<integer>]
[loop-timeout=<integer>] [msgwait-timeout=<integer>]
[crashdump-watchdog-timeout=<integer>] [watchdog-device=<device>]
# For disk-less SBD
crm sbd configure show [sysconfig|property]
-crm sbd configure [watchdog-timeout=<integer>] [watchdog-device=<device>]
+crm sbd configure [watchdog-timeout=<integer>]
[crashdump-watchdog-timeout=<integer>] [watchdog-device=<device>]
...............
example:
@@ -2189,10 +2195,13 @@
Disable the systemd sbd.service on all cluster nodes,
move the sbd sysconfig to .bak and adjust SBD related cluster properties.
+If `crashdump` is specified, the crashdump related configurations will be
+removed.
Usage:
...............
purge
+purge crashdump
...............
[[cmdhelp.node,Node management]]
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/crmsh-5.0.0+20250403.1442a04a/test/features/sbd_ui.feature
new/crmsh-5.0.0+20250418.bb6c26c2/test/features/sbd_ui.feature
--- old/crmsh-5.0.0+20250403.1442a04a/test/features/sbd_ui.feature
2025-04-03 10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/test/features/sbd_ui.feature
2025-04-18 16:21:27.000000000 +0200
@@ -38,7 +38,7 @@
When Try "crm sbd configure watchdog-timeout=f"
Then Except "ERROR: Invalid timeout value: f"
When Try "crm sbd configure name=testing"
- Then Except "ERROR: Unknown argument: name=testing"
+ Then Except "ERROR: Invalid argument: name=testing"
When Try "crm sbd device add /dev/sda6 /dev/sda6"
Then Expected "Duplicated device path detected" in stderr
When Try "crm sbd device add /dev/sda6 /dev/sda7 /dev/sda8"
@@ -49,6 +49,26 @@
When Run "crm sbd configure watchdog-timeout=30 msgwait-timeout=60" on
"hanode1"
Then Run "crm sbd configure show disk_metadata|grep -E "watchdog.*30""
OK
Then Run "crm sbd configure show disk_metadata|grep -E "msgwait.*60"" OK
+ When Run "crm sbd configure watchdog-timeout=30 msgwait-timeout=60" on
"hanode1"
+ Then Expected "No change in SBD configuration" in stdout
+
+ Scenario: sbd configure for crashdump, disk-based sbd case
+ When Run "crm sbd configure crashdump-watchdog-timeout=60" on "hanode1"
+ Then Run "crm sbd configure show disk_metadata|grep -E "watchdog.*30""
OK
+ Then Run "crm sbd configure show disk_metadata|grep -E "msgwait.*120""
OK
+ Then Run "crm configure show stonith-sbd|grep crashdump=1" OK
+ Then Run "crm sbd configure show sysconfig |grep
SBD_TIMEOUT_ACTION=flush,crashdump" OK
+ Then Run "crm sbd configure show sysconfig |grep "SBD_OPTS=\"-C 60\"""
OK
+ When Run "crm sbd configure crashdump-watchdog-timeout=60" on "hanode1"
+ Then Expected "No change in SBD configuration" in stdout
+ # Purge crashdump
+ When Run "crm sbd purge crashdump" on "hanode1"
+ And Try "crm configure show stonith-sbd|grep crashdump"
+ Then Expected return code is "1"
+ When Try "crm sbd configure show sysconfig |grep
SBD_TIMEOUT_ACTION=flush,crashdump"
+ Then Expected return code is "1"
+ When Try "crm sbd configure show sysconfig |grep "SBD_OPTS=\"-C 60\"""
+ Then Expected return code is "1"
Scenario: sbd device add and remove
# Add a sbd disk
@@ -57,7 +77,7 @@
Then Run "crm sbd configure show sysconfig|grep -E
"SBD_DEVICE=\"/dev/sda5;/dev/sda6\""" OK
Then Run "crm sbd configure show sysconfig|grep -E
"SBD_DEVICE=\"/dev/sda5;/dev/sda6\""" OK on "hanode2"
And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda6'|grep -E "watchdog.*30"" OK
- And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda6'|grep -E "msgwait.*60"" OK
+ And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda6'|grep -E "msgwait.*120"" OK
When Run "crm cluster restart --all" on "hanode1"
And Wait for DC
# Remove a sbd disk
@@ -71,7 +91,7 @@
Then Run "crm sbd configure show sysconfig|grep -E
"SBD_DEVICE=\"/dev/sda6;/dev/sda7\""" OK
Then Run "crm sbd configure show sysconfig|grep -E
"SBD_DEVICE=\"/dev/sda6;/dev/sda7\""" OK on "hanode2"
And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda7'|grep -E "watchdog.*30"" OK
- And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda7'|grep -E "msgwait.*60"" OK
+ And Run "crm sbd configure show disk_metadata |grep -A 8
'/dev/sda7'|grep -E "msgwait.*120"" OK
When Run "crm cluster restart --all" on "hanode1"
And Wait for DC
# Purge sbd from cluster
@@ -93,6 +113,20 @@
# Shoud not has any sbd device configured
When Try "crm sbd configure show sysconfig|grep -E "SBD_DEVICE=.+""
Then Expected return code is "1"
+ # enable crashdump
+ Then Run "crm sbd configure show sysconfig |grep
SBD_WATCHDOG_TIMEOUT=15" OK
+ When Run "crm sbd configure crashdump-watchdog-timeout=60" on "hanode1"
+ Then Run "crm sbd configure show sysconfig |grep
SBD_TIMEOUT_ACTION=flush,crashdump" OK
+ Then Run "crm sbd configure show sysconfig |grep "SBD_OPTS=\"-C 60
-Z\""" OK
+ Then Run "crm sbd configure show property |grep
stonith-watchdog-timeout=75" OK
+ When Run "crm sbd configure crashdump-watchdog-timeout=60" on "hanode1"
+ Then Expected "No change in SBD configuration" in stdout
+ # Purge crashdump
+ When Run "crm sbd purge crashdump" on "hanode1"
+ When Try "crm sbd configure show sysconfig |grep
SBD_TIMEOUT_ACTION=flush,crashdump"
+ Then Expected return code is "1"
+ When Try "crm sbd configure show sysconfig |grep "SBD_OPTS=\"-C 60
-Z\"""
+ Then Expected return code is "1"
# Purge sbd from cluster
When Run "crm sbd purge" on "hanode1"
And Run "crm cluster restart --all" on "hanode1"
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_sbd.py
new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_sbd.py
--- old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_sbd.py
2025-04-03 10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_sbd.py
2025-04-18 16:21:27.000000000 +0200
@@ -97,6 +97,18 @@
result = SBDUtils.get_sbd_device_from_config()
self.assertEqual(result, ['/dev/sbd_device',
'/dev/another_sbd_device'])
+ @patch('crmsh.sbd.SBDUtils.get_sbd_value_from_config')
+ def test_get_crashdump_watchdog_timeout_none(self,
mock_get_sbd_value_from_config):
+ mock_get_sbd_value_from_config.return_value = None
+ result = SBDUtils.get_crashdump_watchdog_timeout()
+ self.assertIsNone(result)
+
+ @patch('crmsh.sbd.SBDUtils.get_sbd_value_from_config')
+ def test_get_crashdump_watchdog_timeout(self,
mock_get_sbd_value_from_config):
+ mock_get_sbd_value_from_config.return_value = "-C 60 -Z"
+ result = SBDUtils.get_crashdump_watchdog_timeout()
+ self.assertEqual(result, 60)
+
@patch('crmsh.sbd.SBDUtils.get_sbd_device_from_config')
@patch('crmsh.service_manager.ServiceManager.service_is_active')
def test_is_using_diskless_sbd(self, mock_service_is_active,
mock_get_sbd_device_from_config):
@@ -401,7 +413,11 @@
mock_CrmMonXmlParser.return_value.is_any_resource_running.return_value
= True
SBDManager.restart_cluster_if_possible()
mock_ServiceManager.return_value.service_is_active.assert_called_once_with(constants.PCMK_SERVICE)
- mock_logger_warning.assert_called_once_with("Resource is running, need
to restart cluster service manually on each node")
+ mock_logger_warning.assert_has_calls([
+ call("Resource is running, need to restart cluster service
manually on each node"),
+ call("Or, run with `crm -F` or `--force` option, the `sbd`
subcommand will leverage maintenance mode for any changes that require
restarting sbd.service"),
+ call("Understand risks that running RA has no cluster protection
while the cluster is in maintenance mode and restarting")
+ ])
@patch('crmsh.bootstrap.restart_cluster')
@patch('logging.Logger.warning')
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_ui_sbd.py
new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_ui_sbd.py
--- old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_ui_sbd.py
2025-04-03 10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_ui_sbd.py
2025-04-18 16:21:27.000000000 +0200
@@ -90,7 +90,7 @@
mock_ServiceManager.return_value.service_is_active.side_effect =
[True, True]
mock_is_using_disk_based_sbd.return_value = False
mock_is_using_diskless_sbd.return_value = True
- self.assertEqual(ui_sbd.sbd_configure_completer(["configure", ""]),
["show", "watchdog-timeout=", "watchdog-device="])
+ self.assertEqual(ui_sbd.sbd_configure_completer(["configure", ""]),
["show", "watchdog-timeout=", "crashdump-watchdog-timeout=",
"watchdog-device="])
mock_ServiceManager.return_value.service_is_active.assert_has_calls([
mock.call(constants.PCMK_SERVICE),
mock.call(constants.SBD_SERVICE)
@@ -115,7 +115,7 @@
mock_get_sbd_device_from_config.return_value = ["/dev/sda1"]
mock_get_watchdog_device_from_sbd_config.return_value =
"/dev/watchdog0"
mock_get_sbd_watchdog_timeout.return_value = 10
- mock_get_sbd_device_metadata.return_value = {"watchdog": 10,
"msgwait": 20}
+ mock_get_sbd_device_metadata.return_value = {"watchdog": 10,
"allocate": 5, "loop": 5, "msgwait": 20}
self.sbd_instance_diskbased = ui_sbd.SBD()
self.sbd_instance_diskbased._load_attributes()
@@ -198,14 +198,14 @@
def test_do_configure_no_service(self):
self.sbd_instance_diskbased.service_is_active =
mock.Mock(return_value=False)
- res = self.sbd_instance_diskbased.do_configure(mock.Mock())
+ res = self.sbd_instance_diskbased.do_configure(mock.Mock(), "test")
self.assertFalse(res)
@mock.patch('crmsh.sbd.SBDTimeout.get_sbd_systemd_start_timeout')
@mock.patch('logging.Logger.info')
@mock.patch('builtins.print')
def test_show_property(self, mock_print, mock_logger_info,
mock_get_sbd_systemd_start_timeout):
- data = """property cib-bootstrap-options: \
+ data1 = """property cib-bootstrap-options: \
have-watchdog=true \
dc-version="2.1.7+20240711.239cba384-1.1-2.1.7+20240711.239cba384" \
cluster-infrastructure=corosync \
@@ -214,18 +214,21 @@
stonith-timeout=83 \
priority-fencing-delay=60
"""
- self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error =
mock.Mock(return_value=data)
+ data2 = "fence_sbd parameters"
+ self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error =
mock.Mock(side_effect=[data1, data2])
mock_get_sbd_systemd_start_timeout.return_value = 10
self.sbd_instance_diskbased._show_property()
mock_logger_info.assert_has_calls([
mock.call("crm sbd configure show property"),
+ mock.call("%s", "crm configure show related:fence_sbd"),
mock.call("%s", sbd.SBDTimeout.SHOW_SBD_START_TIMEOUT_CMD)
])
mock_print.assert_has_calls([
mock.call("have-watchdog=true"),
mock.call("stonith-enabled=true"),
mock.call("stonith-timeout=83"),
- mock.call("priority-fencing-delay=60"),
+ mock.call(),
+ mock.call("fence_sbd parameters"),
mock.call(),
mock.call(f"TimeoutStartUSec=10")
])
@@ -255,6 +258,19 @@
self.sbd_instance_diskbased._configure_show(["show", "property"])
self.sbd_instance_diskbased._show_property.assert_called_once()
+ def test_parse_re(self):
+ test_data = [
+ ("watchdog-timeout=30", ("watchdog", "timeout", "30")),
+ ("crashdump-watchdog-timeout=120", ("crashdump-watchdog",
"timeout", "120")),
+ ("watchdog-device=/dev/watchdog", ("watchdog", "device",
"/dev/watchdog")),
+ ("loop-timeout=5", ("loop", "timeout", "5")),
+ ("msgwait-timeout=10", ("msgwait", "timeout", "10")),
+ ]
+ for input_str, expected in test_data:
+ match = ui_sbd.SBD.PARSE_RE.match(input_str)
+ self.assertIsNotNone(match)
+ self.assertEqual(match.groups(), expected)
+
@mock.patch('crmsh.ui_sbd.SBD._show_sysconfig')
@mock.patch('builtins.print')
def test_configure_show(self, mock_print, mock_show_sysconfig):
@@ -273,11 +289,6 @@
self.sbd_instance_diskbased._parse_args(["watchdog-timeout=xxx"])
self.assertEqual(str(e.exception), "Invalid timeout value: xxx")
- def test_parse_args_unknown_arg(self):
- with self.assertRaises(ui_sbd.SBD.SyntaxError) as e:
- self.sbd_instance_diskbased._parse_args(["name=xin"])
- self.assertEqual(str(e.exception), "Unknown argument: name=xin")
-
@mock.patch('logging.Logger.debug')
@mock.patch('crmsh.watchdog.Watchdog.get_watchdog_device')
def test_parse_args(self, mock_get_watchdog_device, mock_logger_debug):
@@ -299,60 +310,139 @@
res = ui_sbd.SBD._adjust_timeout_dict(timeout_dict)
self.assertEqual(res, {"watchdog": 5, "msgwait": 10})
+ @mock.patch('logging.Logger.error')
+ @mock.patch('crmsh.cibquery.get_primitives_with_ra')
+ @mock.patch('crmsh.cibquery.ResourceAgent')
+ @mock.patch('crmsh.xmlutil.text2elem')
+ def test_set_crashdump_option_exception(self, mock_text2elem,
mock_ResourceAgent, mock_get_primitives_with_ra, mock_logger_error):
+ self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error =
mock.Mock(return_value="<dummy></dummy>")
+ mock_text2elem.return_value = "dummy"
+ mock_ra_instance = mock.Mock()
+ mock_ResourceAgent.return_value = mock_ra_instance
+ mock_get_primitives_with_ra.return_value = []
+
+ with self.assertRaises(ui_sbd.SBD.MissingRequiredException):
+ self.sbd_instance_diskbased._set_crashdump_option()
+
+
self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error.assert_called_once_with("crm
configure show xml")
+ mock_logger_error.assert_called_once_with("No fence_sbd resource
found")
+
+ @mock.patch('logging.Logger.info')
+ @mock.patch('crmsh.utils.is_boolean_false')
+ @mock.patch('crmsh.cibquery.get_parameter_value')
+ @mock.patch('crmsh.cibquery.get_primitives_with_ra')
+ @mock.patch('crmsh.cibquery.ResourceAgent')
+ @mock.patch('crmsh.xmlutil.text2elem')
+ def test_set_crashdump_option(self, mock_text2elem, mock_ResourceAgent,
mock_get_primitives_with_ra, mock_get_parameter_value, mock_is_boolean_false,
mock_logger_info):
+ self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error =
mock.Mock(side_effect=["<dummy></dummy>", ""])
+ mock_text2elem.return_value = "dummy"
+ mock_ra_instance = mock.Mock()
+ mock_ResourceAgent.return_value = mock_ra_instance
+ mock_get_primitives_with_ra.return_value = ["fence_sbd"]
+ mock_get_parameter_value.return_value = None
+ mock_is_boolean_false.return_value = True
+
+ self.sbd_instance_diskbased._set_crashdump_option()
+ mock_logger_info.assert_called_once_with("Set crashdump option for
fence_sbd resource")
+
+ @mock.patch('logging.Logger.info')
+ @mock.patch('crmsh.utils.is_boolean_false')
+ @mock.patch('crmsh.cibquery.get_parameter_value')
+ @mock.patch('crmsh.cibquery.get_primitives_with_ra')
+ @mock.patch('crmsh.cibquery.ResourceAgent')
+ @mock.patch('crmsh.xmlutil.text2elem')
+ def test_set_crashdump_option_delete(self, mock_text2elem,
mock_ResourceAgent, mock_get_primitives_with_ra, mock_get_parameter_value,
mock_is_boolean_false, mock_logger_info):
+ self.sbd_instance_diskbased.cluster_shell.get_stdout_or_raise_error =
mock.Mock(side_effect=["<dummy></dummy>", ""])
+ mock_text2elem.return_value = "dummy"
+ mock_ra_instance = mock.Mock()
+ mock_ResourceAgent.return_value = mock_ra_instance
+ mock_get_primitives_with_ra.return_value = ["fence_sbd"]
+ mock_get_parameter_value.return_value = None
+ mock_is_boolean_false.return_value = False
+
+ self.sbd_instance_diskbased._set_crashdump_option(delete=True)
+ mock_logger_info.assert_called_once_with("Delete crashdump option for
fence_sbd resource")
+
+ @mock.patch('logging.Logger.warning')
+ def test_check_kdump_service(self, mock_logger_warning):
+ self.sbd_instance_diskbased.service_manager.service_is_active =
mock.Mock(side_effect=[True, False])
+ self.sbd_instance_diskbased._check_kdump_service()
+ mock_logger_warning.assert_has_calls([
+ mock.call("Kdump service is not active on %s", "node2"),
+ mock.call("Kdump service is required for crashdump")
+ ])
+
+ def test_should_configure_crashdump_no_set(self):
+ self.sbd_instance_diskbased.crashdump_watchdog_timeout_from_config =
None
+ res = self.sbd_instance_diskbased._should_configure_crashdump(None,
None)
+ self.assertFalse(res)
+
+ def test_should_configure_crashdump(self):
+ self.sbd_instance_diskbased.crashdump_watchdog_timeout_from_config = 1
+ res = self.sbd_instance_diskbased._should_configure_crashdump(10, None)
+ self.assertTrue(res)
+
@mock.patch("crmsh.ui_sbd.SBD.configure_usage",
new_callable=mock.PropertyMock)
@mock.patch('builtins.print')
@mock.patch('logging.Logger.error')
def test_do_configure_no_args(self, mock_logger_error, mock_print,
mock_configure_usage):
- self.sbd_instance_diskbased.service_is_active =
mock.Mock(side_effect=[True, True])
+ self.sbd_instance_diskbased._load_attributes = mock.Mock()
mock_configure_usage.return_value = "usage data"
res = self.sbd_instance_diskbased.do_configure(mock.Mock())
self.assertFalse(res)
mock_logger_error.assert_called_once_with('%s', "No argument")
mock_print.assert_called_once_with("usage data")
+ @mock.patch('logging.Logger.info')
@mock.patch('crmsh.sbd.SBDManager')
- def test_configure_diskbase(self, mock_SBDManager):
- parameter_dict = {"watchdog": 12, "watchdog-device":
"/dev/watchdog100"}
- self.sbd_instance_diskbased._adjust_timeout_dict =
mock.Mock(return_value=parameter_dict)
+ def test_configure_diskbase(self, mock_SBDManager, mock_logger_info):
+ parameter_dict = {"watchdog": 12, "watchdog-device":
"/dev/watchdog100", "crashdump-watchdog": 12}
+ self.sbd_instance_diskbased._should_configure_crashdump =
mock.Mock(return_value=True)
+ self.sbd_instance_diskbased._check_kdump_service = mock.Mock()
+ self.sbd_instance_diskbased._set_crashdump_option = mock.Mock()
+ self.sbd_instance_diskbased._set_crashdump_in_sysconfig =
mock.Mock(return_value={"SBD_TIMEOUT_ACTION": "flush,crashdump", "SBD_OPTS":
"-C 12"})
mock_SBDManager.return_value.init_and_deploy_sbd = mock.Mock()
self.sbd_instance_diskbased._configure_diskbase(parameter_dict)
mock_SBDManager.assert_called_once_with(
device_list_to_init=self.sbd_instance_diskbased.device_list_from_config,
- timeout_dict={"watchdog": 12, "msgwait": 20, "watchdog-device":
"/dev/watchdog100"},
- update_dict={
- "SBD_WATCHDOG_DEV": "/dev/watchdog100"
- }
+ timeout_dict={'watchdog': 12, 'allocate': 5, 'loop': 5, 'msgwait':
36},
+ update_dict={'SBD_TIMEOUT_ACTION': 'flush,crashdump', 'SBD_OPTS':
'-C 12', 'SBD_WATCHDOG_DEV': '/dev/watchdog100'}
)
mock_SBDManager.return_value.init_and_deploy_sbd.assert_called_once()
+ self.sbd_instance_diskbased._check_kdump_service.assert_called_once()
+ self.sbd_instance_diskbased._set_crashdump_option.assert_called_once()
@mock.patch('logging.Logger.info')
- @mock.patch('crmsh.utils.is_subdict')
@mock.patch('crmsh.sbd.SBDManager')
- def test_configure_diskbase_no_change(self, mock_SBDManager,
mock_is_subdict, mock_logger_info):
- parameter_dict = {"watchdog": 10, "watchdog-device": "/dev/watchdog0"}
- mock_is_subdict.return_value = True
+ def test_configure_diskbase_no_change(self, mock_SBDManager,
mock_logger_info):
+ parameter_dict = {"msgwait": 20, "watchdog": 10, "watchdog-device":
"/dev/watchdog0"}
+ self.sbd_instance_diskbased._should_configure_crashdump =
mock.Mock(return_value=False)
self.sbd_instance_diskbased._configure_diskbase(parameter_dict)
mock_logger_info.assert_called_once_with("No change in SBD
configuration")
+ @mock.patch('logging.Logger.info')
@mock.patch('crmsh.sbd.SBDManager')
- def test_configure_diskless(self, mock_SBDManager):
- parameter_dict = {"watchdog": 12, "watchdog-device":
"/dev/watchdog100"}
- self.sbd_instance_diskless._adjust_timeout_dict =
mock.Mock(return_value=parameter_dict)
+ def test_configure_diskless(self, mock_SBDManager, mock_logger_info):
+ parameter_dict = {"watchdog": 12, "watchdog-device":
"/dev/watchdog100", "crashdump-watchdog": 12}
+ self.sbd_instance_diskless._should_configure_crashdump =
mock.Mock(return_value=True)
+ self.sbd_instance_diskless._check_kdump_service = mock.Mock()
+ self.sbd_instance_diskless._check_kdump_service = mock.Mock()
+ self.sbd_instance_diskless._set_crashdump_in_sysconfig =
mock.Mock(return_value={"SBD_TIMEOUT_ACTION": "flush,crashdump", "SBD_OPTS":
"-C 12 -Z"})
mock_SBDManager.return_value.init_and_deploy_sbd = mock.Mock()
self.sbd_instance_diskless._configure_diskless(parameter_dict)
mock_SBDManager.assert_called_once_with(
- update_dict={
- "SBD_WATCHDOG_DEV": "/dev/watchdog100",
- "SBD_WATCHDOG_TIMEOUT": "12"
- },
+ timeout_dict={'stonith-watchdog': 24},
+ update_dict={'SBD_WATCHDOG_TIMEOUT': '12', 'SBD_WATCHDOG_DEV':
'/dev/watchdog100', 'SBD_TIMEOUT_ACTION': 'flush,crashdump', 'SBD_OPTS': '-C 12
-Z'},
diskless_sbd=True
)
mock_SBDManager.return_value.init_and_deploy_sbd.assert_called_once()
+ self.sbd_instance_diskless._check_kdump_service.assert_called_once()
@mock.patch('logging.Logger.info')
@mock.patch('crmsh.sbd.SBDManager')
def test_configure_diskless_no_change(self, mock_SBDManager,
mock_logger_info):
parameter_dict = {"watchdog": 10, "watchdog-device": "/dev/watchdog0"}
+ self.sbd_instance_diskless._should_configure_crashdump =
mock.Mock(return_value=False)
self.sbd_instance_diskless._configure_diskless(parameter_dict)
mock_logger_info.assert_called_once_with("No change in SBD
configuration")
@@ -447,6 +537,7 @@
def test_do_device_add(self, mock_is_using_disk_based_sbd,
mock_logger_info):
mock_is_using_disk_based_sbd.return_value = True
self.sbd_instance_diskbased.service_is_active =
mock.Mock(return_value=True)
+ self.sbd_instance_diskbased._load_attributes = mock.Mock()
self.sbd_instance_diskbased._device_add = mock.Mock()
res = self.sbd_instance_diskbased.do_device(mock.Mock(), "add",
"/dev/sda2;/dev/sda3")
self.assertTrue(res)
@@ -458,6 +549,7 @@
def test_do_device_remove(self, mock_is_using_disk_based_sbd,
mock_logger_info):
mock_is_using_disk_based_sbd.return_value = True
self.sbd_instance_diskbased.service_is_active =
mock.Mock(return_value=True)
+ self.sbd_instance_diskbased._load_attributes = mock.Mock()
self.sbd_instance_diskbased._device_remove = mock.Mock()
res = self.sbd_instance_diskbased.do_device(mock.Mock(), "remove",
"/dev/sda1")
self.assertTrue(res)
@@ -471,14 +563,12 @@
self.assertFalse(res)
mock_purge_sbd_from_cluster.assert_not_called()
- @mock.patch('crmsh.sbd.SBDManager.restart_cluster_if_possible')
@mock.patch('crmsh.sbd.purge_sbd_from_cluster')
- def test_do_purge(self, mock_purge_sbd_from_cluster,
mock_restart_cluster_if_possible):
+ def test_do_purge(self, mock_purge_sbd_from_cluster):
self.sbd_instance_diskbased.service_is_active =
mock.Mock(return_value=True)
res = self.sbd_instance_diskbased.do_purge(mock.Mock())
self.assertTrue(res)
mock_purge_sbd_from_cluster.assert_called_once()
- mock_restart_cluster_if_possible.assert_called_once()
@mock.patch('crmsh.xmlutil.CrmMonXmlParser')
def test_print_sbd_agent_status(self, mock_CrmMonXmlParser):
diff -urN '--exclude=CVS' '--exclude=.cvsignore' '--exclude=.svn'
'--exclude=.svnignore'
old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_utils.py
new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_utils.py
--- old/crmsh-5.0.0+20250403.1442a04a/test/unittests/test_utils.py
2025-04-03 10:05:11.000000000 +0200
+++ new/crmsh-5.0.0+20250418.bb6c26c2/test/unittests/test_utils.py
2025-04-18 16:21:27.000000000 +0200
@@ -1342,7 +1342,52 @@
mock_error.assert_called_once_with('Operation is denied. The current user
lacks the necessary privilege.')
-def test_is_subdict():
- d1 = {"a": 1, "b": 2}
- d2 = {"a": 1}
- assert utils.is_subdict(d2, d1) is True
+@mock.patch('logging.Logger.warning')
+@mock.patch('crmsh.utils.is_dc_idle')
+def test_leverage_maintenance_mode_skip(mock_idle, mock_warn):
+ config.core.force = True
+ mock_idle.return_value = False
+ with utils.leverage_maintenance_mode() as result:
+ assert result is False
+ mock_warn.assert_called_once_with("Pacemaker state transition is in
progress. Skip restarting cluster in maintenance mode.")
+
+
+@mock.patch('crmsh.utils.delete_property')
+@mock.patch('crmsh.utils.set_property')
+@mock.patch('logging.Logger.info')
+@mock.patch('crmsh.utils.is_dc_idle')
+def test_leverage_maintenance_mode(mock_idle, mock_info, mock_set,
mock_delete):
+ config.core.force = True
+ mock_idle.return_value = True
+ with utils.leverage_maintenance_mode() as result:
+ assert result is True
+ mock_set.assert_called_once_with("maintenance-mode", "true")
+ mock_delete.assert_called_once_with("maintenance-mode")
+
+
+@mock.patch('crmsh.utils.get_dc')
+def test_is_dc_idle_no_dc(mock_dc):
+ mock_dc.return_value = None
+ assert utils.is_dc_idle() is False
+
+
+@mock.patch('logging.Logger.error')
+@mock.patch('crmsh.utils.ShellUtils')
+@mock.patch('crmsh.utils.get_dc')
+def test_is_dc_idle_failed_get_dc_status(mock_dc, mock_shell, mock_error):
+ mock_dc.return_value = "test"
+ mock_shell_inst = mock.Mock()
+ mock_shell.return_value = mock_shell_inst
+ mock_shell_inst.get_stdout_stderr.return_value = (1, None, "error")
+ assert utils.is_dc_idle() is False
+ mock_error.assert_called_once_with("Failed to get DC status: %s", "error")
+
+
+@mock.patch('crmsh.utils.ShellUtils')
+@mock.patch('crmsh.utils.get_dc')
+def test_is_dc_idle(mock_dc, mock_shell):
+ mock_dc.return_value = "test"
+ mock_shell_inst = mock.Mock()
+ mock_shell.return_value = mock_shell_inst
+ mock_shell_inst.get_stdout_stderr.return_value = (0, "in S_IDLE: ok", None)
+ assert utils.is_dc_idle() is True
