Replacing a Failed Cluster Lock Disk in MC-Serviceguard

Step1: Replace the defective disk.

Step 2: Execute vgcfgrestore to rebuild the LVM reserved areas on this disk.

Note: Even though the lock information is stored in the LVM reserved area, the vgcfgrestore command does not reinitialize the lock information. The lock information is stored in the bad block relocation table. There wouldn’t be any reason for vgcfgrestore to replace the BBRA, since it’s likely we replaced the disk anyhow.

# vgcfgrestore -n vglock /dev/rdsk/cXtYdZ              (if a vgcfgbackup was performed after the cluster lock disk was configured into the cluster)

Step 3: If the disk also has logical volumes, follow the standard procedures to either rebuild them or sync them with the mirror.

Step 4: Serviceguard checks the lock disk on an hourly basis. After the vgcfgrestore command, review the syslog file of an active cluster node for not more than one hour and look for a message that the lock disk is healthy again.

Note: If the vgcfgbackup command was not done on the lock disk volume group after the cluster was formed, the following steps must take place to restore the cluster lock.

·         Stop the Cluster

                # cmhaltcl –f

·         Execute the cmapplyconf command on the cluster config file.

                # cmapplyconf –C <cluster-ascii>

·         Restart the cluster.

                # cmruncl

·         Backup the volume group that contains the cluster lock using vgcfgbackup.

            # vgcfgbackup <VG>