[Pacemaker] cmrd problem when building from source

[Andrew Beekhof]

On Mon, Jan 30, 2012 at 2:19 PM, David Vossel <dvossel at redhat.com> wrote:
> Howdy,
>
> I'm interested in contributing to pacemaker, but I am having a little trouble getting pacemaker going from source. Everything works great for me using the FC16 packages though :)
>
> Right now I'm building both corosync and pacemaker from the current head of their git master branches. Maybe this isn't even recommended for development purposes and I should be taking a different approach... If that's the case stop me here and we'll start with that.

Normally thats fine, however the corosync quorum API only just arrived
and at the same time they dropped support for plugins (which pacemaker
had been using).
So things are a little fragile in master at the moment.

I've been working through some of the issues lately, you might have
more joy with my github repo.

>
> The problem I'm having is occurring in the crmd process. For some reason it appears to be in some unrecoverable state on initialization that results in it constantly shutting down and getting brought back up.  Eventually pacemaker stops re-launching crmd and it doesn't appear in my process list anymore.
>
> Below is the log messages pertaining to the crmd process terminating.  The "We're not part of this cluster anymore" seems to be where the problem starts.  Does anyone know what would cause something like this to occur?
>
> --- LOGS
> Jan 28 16:05:52 pcmk-2 crmd[28234]:   notice: init_quorum_connection: Quorum acquired
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: do_ha_control: Connected to the cluster
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: do_started: Delaying start, no membership data (0000000000100000)
> Jan 28 16:05:52 pcmk-2 crmd[28234]:   notice: crmd_peer_update: Status update: Client pcmk-2/crmd now has status [online] (DC=)
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: crm_update_peer: Node pcmk-2: id=343582912 state=unknown addr=(null) votes=0 born=0 seen=0 proc=00000000000000000000000000111312 (new)
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: pcmk_quorum_notification: Membership 0: quorum retained (0)
> Jan 28 16:05:52 pcmk-2 crmd[28234]:    error: check_dead_member: We're not part of the cluster anymore
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: do_state_transition: State transition S_STARTING -> S_RECOVERY [ input=I_ERROR cause=C_FSA_INTERNAL origin=check_dead_member ]
> Jan 28 16:05:52 pcmk-2 crmd[28234]:    error: do_recover: Action A_RECOVER (0000000001000000) not supported
> Jan 28 16:05:52 pcmk-2 crmd[28234]:    error: do_started: Start cancelled... S_RECOVERY
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: do_state_transition: State transition S_RECOVERY -> S_TERMINATE [ input=I_TERMINATE cause=C_FSA_INTERNAL origin=do_recover ]
> Jan 28 16:05:52 pcmk-2 crmd[28234]:     info: do_lrm_control: Disconnected from the LRM
> Jan 28 16:05:52 pcmk-2 crmd[28234]:   notice: terminate_ais_connection: Disconnecting from AI
>
>
> Thanks,
> -- David
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org