Comment 18 for bug 1903745

I should note that one of the reasons I do not suspect the charm is at fault here is that in the bionic sosreport linked in the bug, I do not see the delay that I would expect in a timeout scenario. If the stop timeout were coming into play, I would expect to see a long duration on the restart.

However, we can see the package was upgraded at 06:17:34 :

...
Start-Date: 2020-11-10 06:17:34
Commandline: /usr/bin/unattended-upgrade
Upgrade: pacemaker:amd64 (1.1.18-0ubuntu1.1, 1.1.18-0ubuntu1.3)
End-Date: 2020-11-10 06:17:36
...

Pacemaker service is restarted at 06:17:34 with a SIGTERM:

Nov 10 06:17:34 [51765] juju-caae6f-19-lxd-6 pacemakerd: notice: crm_signal_dispatch: Caught 'Terminated' signal | 15 (invoking handler)

and is restarted at 06:17:35

Nov 10 06:17:36 [41195] juju-caae6f-19-lxd-6 pacemakerd: info: crm_log_init: Changed active directory to /var/lib/pacemaker/cores
Nov 10 06:17:36 [41195] juju-caae6f-19-lxd-6 pacemakerd: info: get_cluster_type: Detected an active 'corosync' cluster
Nov 10 06:17:36 [41195] juju-caae6f-19-lxd-6 pacemakerd: error: sysrq_init: Cannot write to /proc/sys/kernel/sysrq: Permission denied (13)

This doesn't look like its hindered by the timeout configuration.