> > I'm facing the same problem in these days. I'm using Heartbeat and > > the problem is that when the failed node comes back Heartbeat doesn't > > start it automatically as slave. I'll try to put a script that starts > > OpenLDAP automatically as slave at boot time and see what happens. > > Then the master will update automatically (thanks to slurpd) this new > > slave... at last I hope so :-) > > What you could is to add a check in your slapd init script: > if the server holds the heartbeat service address starts slapd as > master, otherwise starts slapd as slave.
I proceed in this manner, but consider when the node1 falls, node2 became master. When node1 "returns" Heartbeat will not execute the slapd script on node1, so I cannot start it as slave, neither node2 can do this (at least as I know). So I must start slapd as slave on node1. -- Diego de Felice
