fsck -p -y /dev/vgprd01dwhdb08/rlvdata hangs during failover

[Adrian Inzunza_1]

Hi there:

We have HP-UX B.11.23 A ia64, for Itanium with a clustered Oracle environment. Right now we're facing some problems during DB packages failover, the fsck -p -y that we use to move the package from the primary node to the secondary node just hangs and it doesn't shows any log or errors so we can take a look at them carefully. I'm a newbie on HP-UX and I don't know where to start. Is there a checklist or similar so I can take a look at and verify that my current configuration is Ok.? This hang issue only happens with one LV, the other databases (6 in total) doesn't presents any problem at all. I hope you can help me.
Regards

[Steven E. Protter · Accepted solution]

Many SG errors are due to problems with the volume group. You need to make sure its properly activated.

I'd recommend manually fixing this logical volume as follows in non-clustered mode:

fsck -F vxfs -o full -y /dev/vgprd01dwhdb08/rlvdata

I'd pull the primary node out of the cluster just to be anal. I'd also have the cluster shut down. It may be unnecessary, but there is something amiss with the logical volume.

Once that is done check that the cluster configuration properly fails over the volume group, allowing the failover node to activate the volume group.

SEP

[Adrian Inzunza_1]

Hi Steve:

I'm trying to follow you with your earlier input about my problem. One other question, if the logical volume is broken are we able to enable the VG on either the primary or the secondary node.? I mean, if is broken surely is fine in the primary node but not in the secondary node. right.? Or maybe the failover process can cause this issue. (just guessing)

As soon as we have a timeframe we'll apply your earlier recommendation.

Best regards

Adrian