Hi all, I am setting up a high available OpenStack control plan. I have three controllers each running HAproxy, pacemaker, keystone, Mariadb Galera, cinder (api +volume), memcahed, rabbitmq and neutron with OVS DVR and VRRP.
Since I am using a single VIP collocated with HAProxy in pacemaker, I needed suggestions / expert advice on : 1. Adding pacemaker resource for keystone. Pacemaker can restart a service if that's down. What would be any other advantage as I am not using a separate VIP for keystone api. 2. I am using ceph as a storage backend for nova (instance disks), cinder (volumes) and glance (images) . I have installed cinder-volume on the controller nodes Do I need to install it on the compute nodes instead. ( knowing that cinder-volume can only work in active/passive mode) Kind regards, Imran _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
