[ClusterLabs] shutdown and restart of complete cluster due to power outage with UPS

[Andrei Borzenkov]

23.01.2019 17:20, Klaus Wenninger пишет:

> 
> And yes dynamic-configuration of two_node should be possible -
> remember that I had to implement that communication with
> corosync into sbd for clusters that are expanded node-by-node
> using pcs.
> 'corosync-cfgtool -R' to reload the config.
> Using 'corosync-cmapctl' to display or directly set keys should
> work as well.
> 

Last time I played with it had some issues - pacemaker did not notice
quorum change until next cluster-recheck-interval expiration. Should be
somewhere in archives.

It is probably acceptable for the described use case, at least with
default 15 minutes recheck interval.