On Tuesday 21 October 2008 16:50:17 Andrew McGill wrote: > On Tuesday 21 October 2008 12:08:04 Adrian Chapela wrote: > > Marian Marinov escribió: > > > On Monday 20 October 2008 18:50:55 Serge Dubrouski wrote: > > >> On Mon, Oct 20, 2008 at 9:46 AM, Marian Marinov <[EMAIL PROTECTED]> > > >> wrote: > > >>> On Monday 20 October 2008 17:50:54 Michael Schwartzkopff wrote: > > >>>> Am Montag, 20. Oktober 2008 16:22 schrieb Marian Marinov: > > >>>>> Hello, > > >>>>> I'm building a cluster which is using drbd as primary/primary with > > >>>>> ocfs2. > > >>>>> > > >>>>> Is there any tutorial on configuring HeartbeatV2 with this setup? > > >>>>> > > >>>>> Regards > > >>>>> Marian Marinov > > >>>> > > >>>> You will need DRBD version 8. > > >>> > > >>> I have setuped the DRBD with OCFS2 already :) > > >>> > > >>> I simply want to setup the Heartbeat configuration for those. > > >> > > >> In a master/master with OCFS2 configuration what would you want > > >> Hearbeat to do? As far as I understand in a such configurations there > > >> are no resources to move. Do you have a third stanby node? > > > > > > If for some reason DRBD disconnects I want to use Heartbeat for auto > > > reconnect. But for now I don't see how this can happen. > > > > With drbd script provided by Heartbeat you couldn't achieve that because > > heartbeat doesn't know if the slave is OK. It only knows there is a > > resource in this node but it doesn't know if it is running. > > > > I am very unhappy with Heartbeat and DRBD... The only way they can work > > is in a Heartbeat v1 using drbddisk. > > On the drbd list it has been mentioned that the OCF script is unsupported. > You may want to raise this on the drbd list. > > For my purposes, I ditched heartbeat, and settled on a cron job which > periodically runs drbdadm connect all, and another which detects > Secondary/Secondary and starts something up. The failover time is 10 > minutes, but the overheads are very low. > > # Both sides: > */5 * * * * root grep -q Unknown /proc/drbd >/dev/null 2>&1 /dev/null > && /sbin/drbdadm connect all >& /dev/null > > > # Preferred primary > 6,11,16,21,26,31,36,41,46,51,56 * * * * root grep -q > Secondary/Secondary /proc/drbd >& /dev/null && /usr/sbin/xm create vmname >
Thanks for all of the replays. For this part of the cluster I decided to do my own failover without heartbeat as it can't work in such situations :( When I finish with the testing of my script I'll put it here in the list for others to use. Marian > >& /dev/null > > > > > Marian > > > > > >>> I'm still reading it, but what I was asking was: Is there a tutorial > > >>> for Heartbeat for such setup? > > >>> > > >>> Cause most of the tutorials are for Primary/Secondary which is not > > >>> what I want to accomplish. > > _______________________________________________ > 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
