[Pacemaker] pacemaker kills corosync

[Andrew Beekhof]

On Thu, Oct 6, 2011 at 2:48 PM, Olivier Nicole
<Olivier.Nicole at cs.ait.ac.th> wrote:
> Thank you Andrew,
>
>> >> Normally pacemaker is loaded by Corosync, ie you just start corosync
>> >> and leave pacemaker stopped and disabled via chkconfig. This is set
>> >> in corosync.conf by this:
>> >>
>> >> service {
>> >>         # Load the Pacemaker Cluster Resource Manager
>> >>         name: pacemaker
>> >>         ver:  0
>> >> }
>> >>
>> >> http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Clusters_from_Scratch/s-configure-corosync.html
>> >
>> > Yes, I followed this HowTo, but it mentions that the directive service
>> > {} should be in /etc/corosync/service.d/pcmk, not in
>> > /etc/corosync/corosync.conf
>>
>> either-or is fine
>>
>> >
>> > And it also mention in the note at the bootm that pacemaker is not
>> > started automatically by corosync, that it will have to be started
>> > later by /etc/init.d/pacemaker start in part 3.2 of the HowTo.
>>
>> Only for "ver: 1", not for "ver: 0"
>
> I am following the HowTo, so I am using "ver: 1" (I just double
> check).
>
> Everything is going exactly as described in the HowTo except one line
> in the log:
>
> Oct  6 10:07:37 cos3 corosync[1314]:   [pcmk  ] info: pcmk_startup: Service: 10
>
> it says Service: 9 in the HowTo.

On RHEL its 10

>
> And when I follow the next step, that is start pacemaker, the problem
> appears.
>
> I don't have the lines:
>
> Feb  8 16:50:38 pcmk-1 pacemakerd: [13990]: info: start_child: Forked child 14022 for process stonith-ng
> Feb  8 16:50:38 pcmk-1 pacemakerd: [13990]: info: start_child: Forked child 14023 for process cib
>
> in the log. ps axf does not show the processes cib, lrmd, etc.
>
> And I notice that the process corosync is dead.

If pacemakerd can't talk to corosync it wont start the daemons.
Is corosync leaving a core file behind?