Hi,

On Fri, May 27, 2011 at 02:04:00PM +0200, Ulrich Windl wrote:
> Hi!
> 
> I have a question due to a problem that causes a node fence:
> When I set up a Volume group on an MD-RAID device, trying to start these 
> resources through pacemaker causes a node fence. I didn't investigate, but I 
> guess the problem is that the resources are already up when the RAs try to 
> start them.
> 
> This berings up the question: Should RA's methods be idempotent? That is: 
> When trying to start a service that is already started, should that be OK, or 
> should it be an error? Likewise: When trying to stop a resource that is 
> already stopped, it it OK, or is it an error?

All RA actions must be idempotent. If not, then the RA has to be
fixed. This is a common problem with some LSB (init) scripts.

> Independent of that I'm having a problem with an NFS server: The exportfs 
> method (or the nfs server) hangs, causing a node fence.

Please post details in a separate thread.

Thanks,

Dejan

> Regards,
> Ulrich
> 
> 
> _______________________________________________
> 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

Reply via email to