On 8/22/20 11:02 AM, Andrei Borzenkov wrote: > 21.08.2020 21:16, Ken Gaillot пишет: >> Previously at shutdown, sbd determined a clean pacemaker shutdown by >> checking whether any resources were running at shutdown. This would >> lead to sbd fencing if pacemaker shut down in maintenance mode with >> resources active. > > What conditions lead to it? I tried to reproduce it, but so far I was > able to set node in maintenance mode and stop pacemaker on it without > any ill effects. This is openSUSE Tumebleweed with sbd at commit > 4b617a1b8f16033ac464cf2d562f0646082c40fa. You shouldn't be able to reproduce this as there was an individual fix for this condition:
commit 824fe834c67fb7bae7feb87607381f9fa8fa2945 Author: Klaus Wenninger <klaus.wennin...@aon.at> Date: Fri Jun 7 19:09:06 2019 +0200 Fix: sbd-pacemaker: assume graceful exit if leftovers are unmanged Point is that sbd now doesn't need to make complicated assumptions anymore if a pacemaker-shutdown was graceful or not as we've introduced a channel via which pacemaker is explicitly telling sbd that it has successfully reached a clean shutdown state. Klaus > _______________________________________________ > Manage your subscription: > https://lists.clusterlabs.org/mailman/listinfo/users > > ClusterLabs home: https://www.clusterlabs.org/ _______________________________________________ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/