According to https://etherpad.openstack.org/p/kilo-crossproject-ha-integration
Nova-scheduler and Cinder-api may cause race conditions then process requests in parallel.
We should put these as a single node type resources at pacemaker and make them migrating between the controller nodes then needed.
According to https:/ /etherpad. openstack. org/p/kilo- crossproject- ha-integration
Nova-scheduler and Cinder-api may cause race conditions then process requests in parallel.
We should put these as a single node type resources at pacemaker and make them migrating between the controller nodes then needed.