I misread and the systemd unit is native, and it already sets the following settings: SendSIGKILL=no TimeoutStopSec=30min TimeoutStartSec=60s
The problem is that most of these failures have been experienced on juju hacluster charm installations, which are overriding these values
$ cat ./systemd/system/pacemaker.service.d/overrides.conf [Service] TimeoutStartSec=180 TimeoutStopSec=60
This was apparently done to fix the following bug: https://bugs.launchpad.net/charms/+source/hacluster/+bug/1654403
FWIW These values are configurable in charm config options. It seems this bug needs to be revisited and/or this bug may need to be retargeted at least in part to charm-hacluster.
I misread and the systemd unit is native, and it already sets the following settings: 30min
SendSIGKILL=no
TimeoutStopSec=
TimeoutStartSec=60s
The problem is that most of these failures have been experienced on juju hacluster charm installations, which are overriding these values
$ cat ./systemd/ system/ pacemaker. service. d/overrides. conf
[Service]
TimeoutStartSec=180
TimeoutStopSec=60
This was apparently done to fix the following bug: /bugs.launchpad .net/charms/ +source/ hacluster/ +bug/1654403
https:/
FWIW These values are configurable in charm config options. It seems this bug needs to be revisited and/or this bug may need to be retargeted at least in part to charm-hacluster.