Release Policy (was: Re: [Linux-HA] 2.1.2 and failover of colocated resources)

[Peter Clapham]

On 20 Sep 2007, at 11:12, Andrew Beekhof wrote:

> On 9/20/07, Peter Clapham <pc7 at sanger.ac.uk> wrote:
>> I've also found that on a busy heartbeat system /var/lib/pengine
>> needs a clean before performing an upgrade on all nodes... but as
>> always YMMV :-)
>
> oh?  what happens if you dont?

Recount from memory of 2.0.7-2.0.8 upgrade so apologies for greyness  
in details...

Basically the ugraded node was taken off line, upgraded (Debian so  
dpkg -i packages) then restarted. At this point the upgraded node  
would attempt to start but failed to join the cluster. Regrettably no  
longer have logs for this, hence I can't recall the precise reason  
for this [apologies for the, therefore, somewhat anecdotal info].

Pretty much following the previously outlined process and restarting  
the system the nodes would achieve a cluster status and DC etc went  
fine BUT no resources were starting up on the upgraded node.  
Repeating the flush and removing the contents of /var/lib/heartbeat/ 
pengine allowed all to come back to life and the resources restarted  
as expected.

Apols again for the lack of logs, but this took place a while back  
and I now perform my testing on clean vmware snapshots and hence none  
of the above now applies :-)

Pete

> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems



Dr Peter Clapham, Informatics System Group
The Wellcome Trust Sanger Institute, Hinxton, Cambridge, CB10 1HH, UK






-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.