I've had no issues with mass power outages nodes just came up and were healed by time I could start VM's. I've done rolling updates before as well.
However I must have forgotten a step as yesterday I shutdown 1 server in my 3-node replicate setup and I have tones of files/shards not healing yet. I've seen it happen in past and I likely just need a relatively short downtime this evening where I shutdown every VM on that volume, then kil the mounts, and let it heal. Do I need to just manually kill the pid of volumes when I plan to shutdown 1 server instead of letting systemd handle it? Centos 7 servers gluster 3.8.3 (yeah I need to update again) *David Gossage* *Carousel Checks Inc. | System Administrator* *Office* 708.613.2284
_______________________________________________ Gluster-users mailing list [email protected] http://lists.gluster.org/mailman/listinfo/gluster-users
