Re: Correct way to put a MySQL replicated environment safely into maintenance mode?

[Alex Walker]

Thanks for the reply. Given that the master/all nodes were going to be to be offline concurrently for maintenance, I wasn't expecting there to be a huge backlog to catch up on, on the replication front, so that wasn't really a notable consideration.

For anybody else's future reference, what I did in the end was take all 3 nodes offline, restart them with skip-networking enabled, then disabled binary logging on the master (set session sql_log_bin = off;) and ran the optimise table statements on all 3 nodes concurrently, before bringing them back online again normally.

This way I was able to minimise time taken, and without breaking replication.

Maintenance is now completed successfully.