Documentation : Geo Document VM snapshotting + Planned Failover
We have no documentation steps for planned failover or simulated failover with VM snapshotting to help restore secondary to it's previous state.
This can be used as a way rehearsal DR situation, where you isolate a secondary node, promote it as primary (in a split brain situation) make whatever tests you want and want to rollback to the previous state using VM snapshot.
Most of our documentation cover what is needed but there are few extra steps to make sure restoring the snapshot will not get the machine in an inconsistent state.
See discussion here for more context: https://gitlab.slack.com/archives/C32LCGC1H/p1531235415000128
Edited by Rachel Nienaber