HP Serviceguard for Linux Base Version 12.00.00 Release Notes, May 2014

Verify that the new OS supports the current cluster hardware configuration and drivers (network
interfaces, bonding driver, and shared disk storage).
NOTE: If there is no HP bonding driver for the Linux version you are using, use the bonding
driver supplied with the Linux distribution.
Do a full back up on each node.
NOTE: HP recommends you to use the rolling upgrade process which:
Helps you upgrade to the latest software version.
Preserves the current OS and cluster configuration.
Keeps running your mission-critical applications.
Before you upgrade, ensure that you read the “Limitations of rolling upgrades and complete the
entire upgrade process before you can use any Serviceguard commands.
CAUTION: If a failure occurs on one node while you are upgrading another, packages and
the applications they contain may not be able to fail over to the node being upgraded.
7.2 Limitations of rolling upgrades
During rolling upgrade, you must issue Serviceguard commands (other than cmrunnode and
cmhaltnode) on nodes that have been upgraded to latest revision of Serviceguard software.
Issue of commands on yet to be upgraded nodes in the cluster will result in failure or inconsistent
execution.
You cannot modify the cluster or package configuration until the upgrade is complete. You
cannot modify the hardware configuration—including the cluster’s network
configuration—during rolling upgrade. This means that you must upgrade all nodes to the
new release before you can modify the configuration file and copy it to all nodes.
None of the features of the newer release of Serviceguard are allowed until all nodes have
been upgraded.
Binary configuration files may be incompatible between releases of Serviceguard. Do not
manually copy configuration files between nodes. The Serviceguard binary configuration file
(cmclconfig) must be in the correct directory or the cluster will not behave properly.
Within a Serviceguard cluster, no more than two versions of Serviceguard can be running
while the rolling upgrade is in progress.
All nodes must be running the same releases of Linux and Serviceguard before the upgrade.
Rolling upgrades are not intended as a means of using mixed releases of Serviceguard or
Linux within the cluster. It is highly recommended that you upgrade all cluster nodes as quickly
as possible to the new release level.
You cannot delete Serviceguard software (via rpm -e) from a node while the cluster is in the
process of rolling upgrade.
This procedure depends on the upgrade or re-install keeping the same device naming
convention and general system configuration. It is possible for devices to change names or
be changed in the scan order in a way that cannot be corrected. If this happens, the cluster
will need to be re-created rather than to be upgraded.
Sgmgr does not detect the rolling upgrade status of that cluster, hence you must use sgmgr
after rolling upgrade is complete. Sgmgr assumes that the cluster is not under rolling upgrade.
7 Rolling software upgrade 15