On Sat, Jan 28, 2012 at 5:45 AM, Peter Scott <pe...@psdt.com> wrote: > On 1/26/2012 8:12 PM, gustavo panizzo <gfa> wrote: >>> >>> crm_mon says it is starting mysqld on mysql01. >>> (I think it actually stopped and restarted mysqld on mysql01 but >>> haven't been able to verify yet. Seems gratuitous, but I can live >>> with that though.) >> >> do not start clustered services from init >> >> this is not pacemaker specific, any cluster vendor will tell you the >> same >> > Okay, but I can't see how that solves my problem, and testing shows it > doesn't. > > mysqld 0:off 1:off 2:off 3:off 4:off 5:off 6:off > > Now when I reboot both machines pacemaker starts mysqld on the preferred > node and never starts it at all on the other node. That's no good because > if the preferred node ever goes down the other node won't have an up-to-date > copy of the database replicated from it. > > We can't be the first people who want to have transparent failover of a > database, and that means that there must be another mysql server for the > first one to replicate all changes to. Is pacemaker simply the wrong tool > to manage this with? Because pacemaker seems to require that only one node > should be running its managed service at a time. Or have I just failed to > understand how to configure it to act differently? Thanks.
You need to look for the master/slave resource type. This might get you started: http://www.mysqlperformanceblog.com/2011/11/29/percona-replication-manager-a-solution-for-mysql-high-availability-with-replication-using-pacemaker/ _______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org