[PR #11750/6c809dd9 backport][stable-12] pacemaker: fix race condition on resource creation (#11877)

pacemaker: fix race condition on resource creation (#11750)

* remove pacemaker wait arg and fix race condition

* fix up pacemaker resource and stonith polling

* add changelog for pacemaker timeout bug

* remove env from test case and fix changelog file name

* Update changelogs/fragments/11750-pacemaker-wait-race-condition.yml



---------


(cherry picked from commit 6c809dd9db)

Co-authored-by: munchtoast <45038532+munchtoast@users.noreply.github.com>
Co-authored-by: Felix Fontein <felix@fontein.de>

plugins/module_utils/pacemaker.py

@@ -5,6 +5,7 @@
 from __future__ import annotations
 
 import re
+import time
 import typing as t
 
 from ansible_collections.community.general.plugins.module_utils.cmd_runner import CmdRunner, cmd_runner_fmt
@@ -59,6 +60,22 @@ def get_pacemaker_maintenance_mode(runner: CmdRunner) -> bool:
         return bool(maintenance_mode_output)
 
 
+def wait_for_resource(runner: CmdRunner, cli_noun: str, name: str, wait: int, sleep_interval: int = 5) -> None:
+    """Poll ``pcs <cli_noun> status <name>`` until the resource reports Started or the wait budget expires.
+
+    Raises an exception if the resource does not reach the Started state within *wait* seconds.
+    """
+    deadline = time.monotonic() + wait
+    while True:
+        with runner("cli_action state name") as ctx:
+            rc, out, err = ctx.run(cli_action=cli_noun, state="status")
+        if out and "Started" in out:
+            return
+        if time.monotonic() >= deadline:
+            raise Exception(f"Timed out waiting {wait}s for {cli_noun} resource '{name}' to start")
+        time.sleep(sleep_interval)
+
+
 def pacemaker_runner(module: AnsibleModule, **kwargs) -> CmdRunner:
     runner_command = ["pcs"]
     runner = CmdRunner(

plugins/modules/pacemaker_resource.py

@@ -149,6 +149,7 @@ from ansible_collections.community.general.plugins.module_utils.module_helper im
 from ansible_collections.community.general.plugins.module_utils.pacemaker import (
     get_pacemaker_maintenance_mode,
     pacemaker_runner,
+    wait_for_resource,
 )
 
 
@@ -237,7 +238,7 @@ class PacemakerResource(StateModuleHelper):
     def state_present(self):
         with self.runner(
             "cli_action state name resource_type resource_option resource_operation resource_meta resource_argument "
-            "resource_clone_ids resource_clone_meta wait",
+            "resource_clone_ids resource_clone_meta",
             output_process=self._process_command_output(
                 not get_pacemaker_maintenance_mode(self.runner), "already exists"
             ),
@@ -247,10 +248,12 @@ class PacemakerResource(StateModuleHelper):
                 cli_action="resource",
                 resource_clone_ids=self.fmt_as_stack_argument(self.module.params["resource_clone_ids"], "clone"),
             )
+        if not self.module.check_mode and self.vars.wait and not get_pacemaker_maintenance_mode(self.runner):
+            wait_for_resource(self.runner, "resource", self.vars.name, self.vars.wait)
 
     def state_cloned(self):
         with self.runner(
-            "cli_action state name resource_clone_ids resource_clone_meta wait",
+            "cli_action state name resource_clone_ids resource_clone_meta",
             output_process=self._process_command_output(
                 not get_pacemaker_maintenance_mode(self.runner), "already a clone resource"
             ),
@@ -260,6 +263,8 @@ class PacemakerResource(StateModuleHelper):
                 cli_action="resource",
                 resource_clone_meta=self.fmt_as_stack_argument(self.module.params["resource_clone_meta"], "meta"),
             )
+        if not self.module.check_mode and self.vars.wait and not get_pacemaker_maintenance_mode(self.runner):
+            wait_for_resource(self.runner, "resource", self.vars.name, self.vars.wait)
 
     def state_enabled(self):
         with self.runner(

plugins/modules/pacemaker_stonith.py

@@ -125,7 +125,7 @@ value:
 """
 
 from ansible_collections.community.general.plugins.module_utils.module_helper import StateModuleHelper
-from ansible_collections.community.general.plugins.module_utils.pacemaker import pacemaker_runner
+from ansible_collections.community.general.plugins.module_utils.pacemaker import pacemaker_runner, wait_for_resource
 
 
 class PacemakerStonith(StateModuleHelper):
@@ -206,7 +206,7 @@ class PacemakerStonith(StateModuleHelper):
 
     def state_present(self):
         with self.runner(
-            "cli_action state name resource_type resource_option resource_operation resource_meta resource_argument agent_validation wait",
+            "cli_action state name resource_type resource_option resource_operation resource_meta resource_argument agent_validation",
             output_process=self._process_command_output(True, "already exists"),
             check_mode_skip=True,
         ) as ctx:
@@ -218,6 +218,8 @@ class PacemakerStonith(StateModuleHelper):
                 resource_meta=self.vars.stonith_metas,
                 resource_argument=self.vars.stonith_argument,
             )
+        if not self.module.check_mode and self.vars.wait:
+            wait_for_resource(self.runner, "stonith", self.vars.name, self.vars.wait)
 
     def state_enabled(self):
         with self.runner(