[Linux-cluster] Priority in failover domains sometimes not honoured

[Lon Hohberger]

On Fri, Aug 03, 2007 at 04:56:19PM +0200, Jos Vos wrote:
> Hi,
> 
> I have the following failover domains:
> 
>       <failoverdomains>
>               <failoverdomain name="prio_host1" ordered="1" restricted="1">
>                       <failoverdomainnode name="host1" priority="1"/>
>                       <failoverdomainnode name="host2" priority="2"/>
>               </failoverdomain>
>               <failoverdomain name="prio_host2" ordered="1" restricted="1">
>                       <failoverdomainnode name="host1" priority="2"/>
>                       <failoverdomainnode name="host2" priority="1"/>
>               </failoverdomain>
>       </failoverdomains>
> 
> But often a service for, say, "prio_host2" stays running on host1,
> (or vice versa), while both hosts are running fine.  In a "stable"
> situation, that service should be moved to host2, AFAIK (and I
> have seen this happen in some situations).
> 
> This is on RHEL 5.0.
> 
> Is this a bug or...?

It sounds like one.  I think I have a fix for it in 5.1 beta as a side
effect of another bug.  I added a smoothing delay for membership
transitions.

Janne was experiencing the same issue, IIRC.

-- 
Lon Hohberger - Software Engineer - Red Hat, Inc.