[Debian-ha-maintainers] Oops breaks resource failover in RHCS

[Fabio M. Di Nitto]

On 03/11/2010 05:31 PM, Guido Günther wrote:
> Hi Ernesto,
> On Tue, Feb 16, 2010 at 09:07:22PM -0500, Ernesto Rodriguez Reina wrote:
>> Hi, once I wrote you because I had a very very similar problem, and I
>> though it was completed solved. Unfortunately I saw the OOPS again. We
>> have repeted some times and always get the same. Here is my scenario:
>>
>> Node master with nodeid=1;
>> Node spare with nodeid=2;
>> Node slave1 with nodeid=3;
>> Node slave2 with nodeid=4;
>>
>> We shutdown node master. Services are corrected relocated. We turn on
>> node Master and again services are corrected relocated. We then
>> shutdown node master again and then the oops appears but only on node
>> spare, nodes slave1 and slave2 seems to be ok with services running.
>> We tested with to different kernels 2.6.32.8 and 2.6.31.5 (with patch
>> http://git.kernel.org/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=063c4c99630c0b06afad080d2a18bda64172c1a2).
>>
>> We are using RHCS 3.0.4-2 from debian mirror. Any ideas of how to
>> solve this? We are going to test with RHCS 3.0.6-5
> There's 3.0.9 in experimental now. Did you report the issue upstream as
> Fabione suggested?
> Cheers,
>   -- Guido

> https://bugzilla.redhat.com/show_bug.cgi?id=566573

It's not a bug in cluster/corosync tho.

Fabio