To be able to deploy Gluster cluster (HA) using VIP through Pacemaker instead of keepalived. The main justifications are, so far: 1. Pacemaker is fully supported on RHEL6+ 2. Already used by Openstack Foreman installer
Hi James, Could you please provide some information about this one? Thanks, Gilles
(In reply to Gilles Dubreuil from comment #1) > Hi James, > > Could you please provide some information about this one? > > Thanks, > Gilles I'm busy working on other projects at the moment, so this isn't a priority right now. Sorry, Cheers, James
Thanks to ndevos for volunteering to fix this bug ;)
(In reply to James (purpleidea) from comment #3) > Thanks to ndevos for volunteering to fix this bug ;) I think this is actually more something Kaleb is (or should be) interested in.
Feature requests make most sense against the 'mainline' release, there is no ETA for an implementation and requests might get forgotten when filed against a particular version.
We already have a glusterfs-resource-agents subpackage which has a (pacemaker) glusterd heartbeat resource. Although it looks like its resource file is installed in /usr/lib/ocf/resource.d/glusterfs/glusterd instead of /usr/lib/resource.d/heartbeat, where the heartbeat resources are usually installed. What are we really asking for here?
(In reply to Kaleb KEITHLEY from comment #6) > We already have a glusterfs-resource-agents subpackage which has a > (pacemaker) glusterd heartbeat resource. > > Although it looks like its resource file is installed in > /usr/lib/ocf/resource.d/glusterfs/glusterd instead of > /usr/lib/resource.d/heartbeat, where the heartbeat resources are usually > installed. > > What are we really asking for here? Actually his is a bit different... I am using this technique: https://ttboj.wordpress.com/2012/08/23/how-to-avoid-cluster-race-conditions-or-how-to-implement-a-distributed-lock-manager-in-puppet/ in my puppet module. It uses keepalived instead of pacemaker for the VIP it needs. This feature requests porting puppet-gluster to also use pacemaker. Actually if GlusterFS or anything else provided an HA identifier, it could be ported to use that instead... Does this answer the question? Cheers!
We are not actively maintaining puppet-gluster. Hence closing this enhancement request.