rootninja

High Availability

With Samba using a clustered database, you can now export the same shared storage (this is only on the gfs2 file system - I haven’t seen any testing on ocfs2) on multiple nodes in an active/active samba cluster. The storage on clients that mount this export will be available when active failover occurs from one server node to another. Multiple clients can mount the export from multiple nodes, all at the same time.

Hardware Requirements

* cluster nodes (2 or more) connected to shared storage
* client node(s) to run client loads from.

Packages

* kernel with gfs2 and dlm enabled
* Red Hat cluster packages cman/openais/corosync/lvm2-cluster
* Latest samba and ctdb packages. (clustered database for samba)

Setup

* Setup the cluster.conf file and start up the cluster with ’service cman start’, and if you’re using clustered lvm with ’server clvmd start’
* mount a gfs2 volume on all nodes (at the same mount point).
* Export the mount point as a samba share (smb.conf in all nodes should be identical).
* Set ctdb configuration.
* start the ctdb service and check ‘ctdb status’ to make sure the ctdb cluster is actually up.
* Done. You can access the same share through all nodes and read/write to them at the same time without a problem.

Tests

* The smbtorture test suite in the samba package has a few scenarios RAW_BENCH_LOCK, RAW_BENCH_OPEN, BENCH_NBENCH
* There’s also the ping_pong test at: ping_pong doc
* Any other test you can think of with multiple clients (windows users mounting the samba share) that stresses ctdb/gfs2.