carta
ctdb_mutex_ceph_rados_helper(7)

ctdb_mutex_ceph_rados_helper7

Ceph rados cluster mutex helper

Overview

ctdb_mutex_ceph_rados_helper can be used as a cluster lock provider for CTDB. When configured, split brain avoidance during CTDB recovery will be handled using locks against an object located in a Ceph RADOS pool. To enable this functionality, include the following line in the [cluster] section of ctdb.conf(5):

cluster lock = !ctdb_mutex_ceph_rados_helper [Cluster] [User] [Pool] [Object] [Timeout] [-n Namespace]
Cluster: Ceph cluster name (e.g. ceph)
User: Ceph cluster user name (e.g. client.admin)
Pool: Ceph RADOS pool name
Object: Ceph RADOS object name
Timeout: Ceph RADOS lock duration in seconds (optional)
Namespace: Ceph RADOS pool namespace (optional)

The Ceph cluster Cluster must be up and running, with a configuration, and keyring file for User located in a librados default search path (e.g. /etc/ceph/). Pool must already exist.

For informational purposes, ctdb_mutex_ceph_rados_helper will also register the cluster lock holder in Ceph Manager's service map.

History

  • Debian 13 (Trixie): the text shown here
  • Debian 12 (Bookworm)
  • Debian 11 (Bullseye)

Author

This documentation was written by David Disseldorp

From ctdb 2:4.22.11+dfsg-0+deb13u1 in Debian 13 (Trixie). Parsed by carta from the page's man source.