Hi all,

I would like to be able to recover automatically the VMS when a compute node dies as described here: http://blog.clusterlabs.org/blog/2015/openstack-ha-compute/

I've checked that I have pacemaker_remote.service and NovaCompute/NovaEvacuate pacemaker resources on my compute nodes, but it's doesn't seem to be configured/running:

[root@overcloud-novacompute-0 openstack]# systemctl list-unit-files | grep pacemaker

pacemaker.service disabled

pacemaker_remote.service disabled

[root@overcloud-novacompute-0 openstack]# pcs status

Error: cluster is not currently running on this node

Is there a way to activate this on stack deployment? Or do I have to customize it?

Thanks,

Pedro Sousa