On 10/19/07, David Teigland <[EMAIL PROTECTED]> wrote: > > On Thu, Oct 18, 2007 at 08:05:02PM +0200, [EMAIL PROTECTED] wrote: > > > So, without qdisk my cluster just behaves exactly as I ever wanted! So > > is qdisk evil? However, please, be so kind to explain this strange > > behavor to me, that's really a _MYSTERY_: Isn't the qdisk a tie-breaker > > useful especially in two node clusters and if one needs to decide > > cluster membership also on the basis of exogenous heuristics (like > > network connectivity as in my case, see RedHat Cluster FAQ)? Shouldn't > > the qdisk allow one to build a more robust cluster against split-brain > > conditions? Why does my cluster behave good only if I avoid using > > qdisk? Have I really no more chance to use a quorum disk in my cluster > > architecture? If it can help, I'd like to tell you that when I start my > > cluster with qdisk enabled, both nodes wait for each other on "Starting > > fencing..." before going on in the boot sequence: no node can boot alone > > while the other one is down. That doesn't happen when I don't use > > qdisk, as you told me. Again, thank you very very much indeed! > > qdisk doesn't work well in RHEL5.0; I'm told it will work in 5.1 > > Dave
Is there a date set for the release? M
-- Linux-cluster mailing list Linux-cluster@redhat.com https://www.redhat.com/mailman/listinfo/linux-cluster
