On Wed, Aug 19, 2009 at 9:51 AM, Alain.Moulle<[email protected]> wrote: > Hi, > Thanks again. And this leads me to another linked question : > if we want to have redundancy for the heartbeat, that means that > we could set 8 lines likewise : > ucast eth0 139.111.12.1 > ucast eth0 139.111.12.2 > ucast eth0 139.111.12.3 > ucast eth0 139.111.12.4 > > ucast eth1 139.222.12.1 > ucast eth1 139.222.12.2 > ucast eth1 139.222.12.3 > ucast eth1 139.222.12.4 > > Right ?
Yes > > But in this case, suppose we lost the eth1 on node 3, so that : > ucast eth1 139.222.12.3 fails from other nodes, does that lead > to stonith of node 3 ? no, because (as you said below) we can still see it via eth0 so there is no need to shoot anyone > or the fact that : > ucast eth0 139.111.12.3 is always ok prevents node3 to be kill by others ? > > Thanks > Alain > > >> Hi, >> > I wonder if we can use ucast for heartbeat in a cluster of more than >> > two-nodes, >> > and if so, in case of for example 4 nodes , I suppose we have to set 4 >> > lines >> > in ha.cf, one line per node in the cluster : >> > ucast eth0 139.111.12.1 >> > ucast eth0 139.111.12.2 >> > ucast eth0 139.111.12.3 >> > ucast eth0 139.111.12.4 >> > but in this case, does it matter to have a "ucast on myself" as ha.cf has >> > to be the same on the 4 nodes ? >> >> >> nope, perfectly fine to do this >> >> >>> > Thanks for your response. >>> > Alain > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
