On 2011-10-13 18:57, Charles Richard wrote: > Hi, > > Have a curiosity question to see if anybody out there is using MySQL with > heartbeat or pacemaker on 2 nodes without using DRBD? And if so, is it an > automatic or a manual failover.
You can use Pacemaker managed MySQL on shared storage, or you can manage MySQL Replication from Pacemaker. Both are possible and supported. > The reason I'm thinking somebody might want to do this is if, as in my case, > management won't invest in Stonith devices so the DRBD/Pacemaker scheme > would work correctly or if the risk of a bad master is too high and people > simply rely on mysql replication with a manual failover when STONITH devices > won't be purchased. If you just wanted manual "failover", you wouldn't be thinking about a cluster. And if you were to run on shared storage, you absolutely, positively MUST use fencing/STONITH. Did I mention you must? Well, you must. > It seems that using DRBD/Pacemaker without stonith devices would bring a > risk of corrupt data which i guess could still be a lesser risk than > potentially having a master with bad data and potential bad backups. I'm not following. What's the difference between "corrupt data" and "bad data" in your scheme? And, also, DRBD is not a backup, just as much as RAID is not backup. Fat-finger DROP TABLE, boom, it's dropped in both replicas. Backup is an issue distinct from storage replication, and an issue you'll have to address separately. Maybe you could start by stating requirements, and then we can make a suggestion which HA option suits your use case best. Cheers, Florian -- Need help with High Availability? http://www.hastexo.com/now _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
