Notes That Apply to Improved Availability

Restart Serviceguard Quorum Server if Quorum Server Node is Re-imaged

You must restart the Serviceguard quorum server if the quorum server node is re-imaged while improved availability is running. Issue the following command on the quorum server node to restart the quorum server and update the /etc/inittab file:

# /opt/hptc/availability/serviceguard/startquorum

Benign Messages

When a node that is providing improved availability of services through Serviceguard is initially configured and booted, a Serviceguard cmcld daemon is started. Part of the processing performed by the cmcld daemon is to verify that the node is correctly configured for availability and to determine whether or not the node should rejoin an already-active Serviceguard cluster.

The following benign error message are generated in the /var/log/nconfig.log output of booting nodes that are configured for availability; you can safely ignore these messages:

info: Executing C11avail nconfigure 
/bin/mknod: `/dev/pidentd': File exists 
/bin/mknod: `/dev/deadman': File exists 
/etc/rc.d/init.d/cmcluster.init: line 107: let: +: syntax 
error: operand expected (error token is "+") 
cmviewconf: Either binary file does not exist, or the user 
doesn't have access to view the cluster configuration. 

You might also see errors about uninitialized variables in AdapterMap.pm at or near lines 711 and 713. You can safely ignore these messages.

Known Limitation if Nagios is Configured for Improved Availability

When Nagios is configured for improved availability, by default, the head node acts as the Nagios master and the other node in the availability set acts as a Nagios monitor.

The known limitation is that you cannot use the /sbin/service command to restart the Nagios monitor on the non-head node in the availability set because of new functionality provided by the /sbin/service script.

Issue the following command to restart the Nagios monitor service on the non-headnode of the availability set:

# /etc/init.d/nagios restart

Running /sbin/service nagios restart on the non-headnode in the availability set causes the nagios master to fail over.

Network Restart Command Negatively Affects Serviceguard

If a node is actively participating in a Serviceguard cluster, the Serviceguard tool manages some HP XC services and their aliases. Because Serviceguard handles relocating these aliases after a node dies, there are no network scripts defined for the aliases. Therefore, when you issue the /sbin/service network restart command on a node in a Serviceguard cluster, the aliases managed by Serviceguard are removed and never re-created. The services will not function correctly and due to the HP XC configuration of Serviceguard, Serviceguard will not detect the failure.

Follow this procedure of the head node to restart the network:

Problem Failing Over Database Package Under Serviceguard

Occasionally, when the head node is running with Serviceguard and the head node becomes unresponsive, the database package fails to start up successfully on the other node in the availability set with the head node.

Serviceguard reports that the database package is running and that any other packages that are dependent upon the database, namely lvs and nagios, will be down. However, the database may not have actually started correctly.

Run the service mysqld status command to verify whether the database is running . If the mysqld service is not running and to recover from this scenario, use the following commands to restart the service under Serviceguard :

To start any services that are dependent on the database service, issue the following commands for each package. These commands enable Serviceguard to restart the database package and start the remaining packages.