|
Written by Frank Schalude
|
|
Wednesday, 18 June 2008 |
|
The current release of HACMP may fail to start (observed with AIX 6.1 and HACMP 5.4.1) if the VGDA time stamp of a shared VG is different on one of the nodes in the cluster.
Even though there is no more file located in /usr/es/sbin/cluster/etc/vg, HACMP still compares the value. The comparison now is made against the value in the ODM (try lsattr -El vgname). Compare this value to the output of lqueryvg -p <any disk from the VG> -T. HACMP may refuse to start even if you have selected to "Automatically correct errors found". The corrective action does not seem to work properly. In this case, perform an importvg -L <VGNAME> <any disk og the VG> as long as the cluster is still offline. This should update your ODM records. |
