I have a 3 node cluster running eventstore on Ubuntu 18.04. When my cronjob which rsyncs the dB to another disk runs. 10% of the time the cluster looses connection long enough to trigger a new leader election.
ionice - c3 rsync… - - bwlimit 5000
So I run rsync with ionice idle and bwlimit of 5mb/s. That did help but it still fails once every 48 runs.
Is there anything else I can do? Its an mostly idle cluster.