From: Adam King <adking@redhat.com>
Date: Wed, 1 Dec 2021 09:01:28 +0000 (-0500)
Subject: mgr/cephadm: don't include agents in CEPHADM_FAILED_DAEMON
X-Git-Tag: v17.1.0~134^2
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=refs%2Fpull%2F44158%2Fhead;p=ceph.git

mgr/cephadm: don't include agents in CEPHADM_FAILED_DAEMON

They already have their own, more strict health warning.
There's very few cases they would show up in failed daemon
health check but not agent down health check and even if
they did it would be temporary. Also, agents marked as down
will automatically (before this change) be marked as failed
even if they don't meet the typical criteria for failed
(systemd status is in error)

Fixes: https://tracker.ceph.com/issues/53448

Signed-off-by: Adam King <adking@redhat.com>
---

diff --git a/src/pybind/mgr/cephadm/module.py b/src/pybind/mgr/cephadm/module.py
index b8cd31931ab..348824f89c2 100644
--- a/src/pybind/mgr/cephadm/module.py
+++ b/src/pybind/mgr/cephadm/module.py
@@ -754,9 +754,10 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule,
     def update_failed_daemon_health_check(self) -> None:
         failed_daemons = []
         for dd in self.cache.get_error_daemons():
-            failed_daemons.append('daemon %s on %s is in %s state' % (
-                dd.name(), dd.hostname, dd.status_desc
-            ))
+            if dd.daemon_type != 'agent':  # agents tracked by CEPHADM_AGENT_DOWN
+                failed_daemons.append('daemon %s on %s is in %s state' % (
+                    dd.name(), dd.hostname, dd.status_desc
+                ))
         self.remove_health_warning('CEPHADM_FAILED_DAEMON')
         if failed_daemons:
             self.set_health_warning('CEPHADM_FAILED_DAEMON', f'{len(failed_daemons)} failed cephadm daemon(s)', len(