|
 |
f312c6 |
From e316840a7e1d2a72e3089ee194334244c959905a Mon Sep 17 00:00:00 2001
|
|
|
f312c6 |
From: Ken Gaillot <kgaillot@redhat.com>
|
|
|
f312c6 |
Date: Wed, 19 Jan 2022 09:53:53 -0600
|
|
|
f312c6 |
Subject: [PATCH] Fix: pacemakerd: tweak systemd unit respawn settings
|
|
|
f312c6 |
|
|
|
f312c6 |
If pacemaker exits immediately after starting, wait 1 second before trying to
|
|
|
f312c6 |
respawn, since the default of 100ms is a bit aggressive for a Pacemaker
|
|
|
f312c6 |
cluster.
|
|
|
f312c6 |
|
|
|
f312c6 |
Also, allow 5 attempts in 25 seconds before giving up.
|
|
|
f312c6 |
---
|
|
|
f312c6 |
daemons/pacemakerd/pacemaker.service.in | 6 ++++++
|
|
|
f312c6 |
1 file changed, 6 insertions(+)
|
|
|
f312c6 |
|
|
|
f312c6 |
diff --git a/daemons/pacemakerd/pacemaker.service.in b/daemons/pacemakerd/pacemaker.service.in
|
|
|
f312c6 |
index 0363a2259c..3fd53d9ffb 100644
|
|
|
f312c6 |
--- a/daemons/pacemakerd/pacemaker.service.in
|
|
|
f312c6 |
+++ b/daemons/pacemakerd/pacemaker.service.in
|
|
|
f312c6 |
@@ -31,6 +31,9 @@ After=rsyslog.service
|
|
|
f312c6 |
After=corosync.service
|
|
|
f312c6 |
Requires=corosync.service
|
|
|
f312c6 |
|
|
|
f312c6 |
+# If Pacemaker respawns repeatedly, give up after this many tries in this time
|
|
|
f312c6 |
+StartLimitBurst=5
|
|
|
f312c6 |
+StartLimitIntervalSec=25s
|
|
|
f312c6 |
|
|
|
f312c6 |
[Install]
|
|
|
f312c6 |
WantedBy=multi-user.target
|
|
|
f312c6 |
@@ -57,6 +60,9 @@ TasksMax=infinity
|
|
|
f312c6 |
# resource. Sending -KILL will just get the node fenced
|
|
|
f312c6 |
SendSIGKILL=no
|
|
|
f312c6 |
|
|
|
f312c6 |
+# Systemd's default of respawning a failed service after 100ms is too aggressive
|
|
|
f312c6 |
+RestartSec=1s
|
|
|
f312c6 |
+
|
|
|
f312c6 |
# If we ever hit the StartLimitInterval/StartLimitBurst limit, and the
|
|
|
f312c6 |
# admin wants to stop the cluster while pacemakerd is not running, it
|
|
|
f312c6 |
# might be a good idea to enable the ExecStopPost directive below.
|
|
|
f312c6 |
--
|
|
|
f312c6 |
2.27.0
|
|
|
f312c6 |
|