On Mon, 2009-07-27 at 16:14 +0200, Lars Ellenberg wrote:
> On Mon, Jul 27, 2009 at 02:49:25PM +0200, Bernd Petrovitsch wrote:
> > Hi all!
> >
> > I think I discovered a bug in heartbeat (or I'm too blind and overlooked
> > it in the documentation).
> > How to trigger:
> > Define the ressource groups in the /etc/ha.d/haresources file so that
> > the first service, say "stop-xxx", after the nodename has/is the same
> > string for both nodes (which is the case if you want something done as
> > the very first step on a failover/failback, e.g. stop various daemons so
> > that they not in the way if filesystems are munmounted) - and not the
> > usual IP address entry.
> >
> > Result: heartbeat (in my case heartbeat-2.1.3-3 from RHEL5/CentOS-5)
> > doesn't really starts up.
> >
> > Debugging the above, it leads to calls of the ResourceManager script
> > with the parameters "givegroup stop-xxx" where actually a nodename is
> > expected.
> > Grepping around reveals that the ip-request script calls
> > "$RSCMGR givegroup $IPADDR" - and that doesn't fit to the above.
>
> "design limitation"
> from http://www.linux-ha.org/haresources
>
> Note: Make sure that the first resource in each resource group is
> unique because the first resource will be used as resource group name.
OK, I'm blind. Or just too lazy to read all the documentation in detail.
> > The obvious solution is to keep an IP address entry on the 1st place
> > (and it works in my case).
> >
> > Or am I doing something seriously wrong?
>
> Yes.
> You are using heartbeat v1, where you should be using pacemaker ;)
I hate to write that but hat wasn't feasible (for not-so-technical
reasons - namely a "simple" migration from
$STALE_AND_VERY_OLD_DISTRIBUTION to $STABLE_DISTRIBUTION and the above
bit me too late ....).
Bernd
--
Firmix Software GmbH http://www.firmix.at/
mobil: +43 664 4416156 fax: +43 1 7890849-55
Embedded Linux Development and Services
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
