Multiple Cinder Nodes

Multiple Cinder Nodes

For multiple Cinder Volume nodes we need to configure coordinated distributed system to replace local file locks.

References:

• Use coordinate distributed systems to replace local file locks
• Coordination configuration options

For example, we can use existing Cinder database as backend for coordinated distributed lock systems:

$ cat /etc/cinder/cinder.conf
[COORDINATION]
backend_url = driver://user:password@host/database?charset=utf8

Available Openstack coordination backend drivers: file, etcd, redis, memcached, mysql, pgsql and zookeeper.