HP Serviceguard for Linux Version A.11.19 Release Notes, October 2009

NOTE: HP recommends you use the rolling upgrade process if possible, because it
keeps your mission-critical applications running; and also that you update the software
(rather than do a cold install) if possible, because updating preserves the current OS
and cluster configuration. But make sure you read the “Special Considerations for
Upgrade to Serviceguard A.11.19” below, as well as the “Limitations ” (page 51) before
you start the upgrade, and keep in mind that the entire upgrade process has to complete
successfully 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 fail over to the node being
upgraded.
Special Considerations for Upgrade to Serviceguard A.11.19
Serviceguard A.11.19 introduces a new cluster manager; see “New Cluster Manager”
(page 12). In a running cluster, this affects node timeout and failover and cluster
re-formation; see the discussion of MEMBER_TIMEOUT under “Cluster Configuration
Parameters” in Chapter 4 of the latest edition of Managing Serviceguard for Linux for
more information. There is also a one-time effect on upgrade to Serviceguard A.11.19.
Upgrade will trigger a cluster membership transition from the old to the new cluster
manager when the last node in the cluster has been upgraded to A.11.19. This transition
can take up to one second, during which time the old cluster manager will shut down
and the new cluster manager will start.
CAUTION: From the time when the old cluster manager is shut down until the new
cluster manager forms its first cluster, a node failure will cause the entire cluster to fail.
HP strongly recommends that you use no Serviceguard commands other than
cmviewcl (1m) until the new cluster manager successfully completes its first cluster
re-formation. In the case of a rolling upgrade, this happens when the cmrunnode for
the last node executes; otherwise (if you have brought down the cluster in order to do
the upgrade) it happens during the initial cmruncl. See “How To Tell when the Cluster
Re-formation Is Complete” for more information.
For an upgrade to Serviceguard A.11.19, this recommendation is additional to the
“Limitations ” below, and supersedes any “Limitation” that is more lenient.
There may be further caveats that apply to specific upgrade paths; before you proceed,
make sure you read the “Announcements” (page 9) and “Compatibility Information
and Installation Requirements” (page 38).
How To Tell when the Cluster Re-formation Is Complete
You can tell when the cluster has re-formed under the new cluster manager by
monitoring syslog on any node. syslog will record that Serviceguard has started
50 Serviceguard for Linux Version A.11.19 Release Notes