OCFS2 Integration with HP Serviceguard for Linux Administrator's Guide, Second Edition, May 2009

1. Upgrade the toolkit on all nodes in the cluster.
# ocfs2mntadmn upgrade node
2. Upgrade all packages that are configured with HP Serviceguard A.11.18.
IMPORTANT: OCFS2 mount point packages that mount or unmount disks configured with
OCFS2 clustered file system have an added level of protection called Persistent Reservation,
available when configured with HP Serviceguard A.11.19. For more information on Persistent
Reservation, see the Managing HP Serviceguard for Linux guide available at http://docs.hp.com ->
High Availability -> Serviceguard for Linux.
Rolling upgrade of mount point packages that were created with HP Serviceguard A.11.18 to
add this new protection feature is currently not supported. A subsequent patch of HP Serviceguard
A.11.19 will include this support. However, OCFS2 mount point packages that were created with
HP Serviceguard A.11.18 will continue to run with HP Serviceguard A.11.19, but with the old
configuration. Currently, if Persistent Reservation is required, then these mount point packages
must be recreated with HP Serviceguard A.11.19 rather than updated from HP Serviceguard
A.11.18.
Pre-configuration Steps
This section elaborates on the pre-configuration steps that you must complete before using OCFS2
in your Serviceguard environment.
This section addresses the following topics:
• “Planning the OCFS2 cluster timeouts” (page 18)
• “Creating the OCFS2 File System” (page 22)
Planning the OCFS2 cluster timeouts
In an environment where the cluster managers of HP Serviceguard for Linux and OCFS2 co-exist
on the same node, it is important to ensure that OCFS2 does not interfere with the recovery
functions of the Serviceguard cluster manager. To ensure that this interference does not occur,
the parameters of the OCFS2 cluster manager must be configured to react slower than the
Serviceguard cluster manager.
Following are the parameters of the OCFS2 cluster manager, O2CB, that need to be configured:
1. O2NET_IDLE_TIMEOUT
2. O2CB_HEARTBEAT_THRESHOLD
3. O2CB_KEEPALIVE_DELAY_MS
4. O2CB_RECONNECT_DELAY_MS
These parameters and the timeout values are stored in the file /etc/sysconfig/o2cb on all
the cluster nodes.
Configuring the O2NET_IDLE_TIMEOUT Parameter
The O2NET_IDLE_TIMEOUT parameter refers to the OCFS2 network idle timeout. This parameter
of OCFS2 specifies the time in milliseconds before a network connection is considered dead.
In the event of a network failure, it is important that the Serviceguard cluster manager realizes
the network failure first and starts the necessary procedures to stabilize the cluster. So, OCFS2
must be configured to realize that there is a network failure only after HP Serviceguard for Linux
has formed a stable cluster. Figure 2-2 describes the time at which HP Serviceguard for Linux
realizes that there is a network failure, and the time when OCFS2 realizes the same.
18 Integrating OCFS2 with HP Serviceguard for Linux