AnsweredAssumed Answered

Restarting many nodes in a cluster

Question asked by snelson on Sep 25, 2014
We're deployed with AWS, and we've been notified that 6 of the 8 nodes in our production cluster will need to be restarted while AWS does maintenance. The six nodes have a six hour window of time in which they will be restarted, but we don't know exactly what time they'll be restarted, and it could happen that all nodes are restarted at exactly the same time. Included in the six nodes are my CLDB master and my one CLDB standby. I've done CLDB failover before, and I'm aware there is a period of time before automatic failover will occur so that the master has time to recover if it's a temporary connection issue. There is no way for us to preemptively restart the instances.

We're planning to just accept the downtime when the CLDB master goes down. So we aren't concerned with availability during the outage, and I'm more concerned about our data. Are there any unpleasant side effects to restarting that many nodes in a short period of time? I assume it's just as safe as restarting Warden?