[CentOS-devel] Corosync/Pacemaker mess after upgrade to 6.4

Tue Jun 11 11:09:43 UTC 2013
France <mailinglists at isg.si>

Hi all,

i've been using Corosync (with Pacemaker) for my production HA iscsi 
service.

Upon upgrading to CentOS 6.4, the upgraded node will not join the 
cluster. It might also be a bug within pacemaker code, but i'm not sure. 
There is also crm command missing (i know i can get in from suse repo). 
I saw the RH's 6.4 technical notes: 
https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/6.4_Technical_Notes/pacemaker.html 
where they suggest we move to cman.

Because the upgrade brakes working cluster with corosync/pacemaker 
setup, i would like to propose, that we add this into to release notes 
for CentOS 6.4.

I'm also thinking about the best way to get fully working cluster back. 
All suggestions are welcome. Please note that we must not stop currently 
running resources.

Is it possible, that there is a bug in new pacemaker (1.1.8-7.el6) and 
it won't work with old pacemaker (1.1.7-6.el6) or are they just 
incompatible be default? (Nodes after upgrade don't see each other.)

I also wonder, will corosync be removed in future RH 6.* releases and 
should i move over to cman?

Regards,
F.