Cluster Lock disks

[Ian Killer_1]

Hi I restored two nodes of a cluster recently. Using tape created by make_tape_recovery. This is a 10.20 cluster. It appears that the recovery didn't write the cluster lock, but the cluster is up and running. SO I'm getting the "cluster lock is missing" message in syslog. I don't have vgcfgbackup files. Does anyone know how to manipulate the bad block relocation table bit that holds the lock?

[melvyn burnard · Accepted solution]

Halt the cluster, activate the clusterlock vg, re-apply the cluster with the cmapplyconf command, or contact your local HP RC and request the cminitlock unsupported utility

[Uday_S_Ankolekar]

Well,

Try this.

Halt the cluster.

De-activate cluster VG

vgchange -a n

Then re-activate then

vgchange -c n

followed by
Cmcheckconf and then cmapplyconf.

-USA..

[Jasja Verbeek_1]

Hi,

Maybe you could try stopping the cluster and then de-activating the shared volumegroup by vgchange -a n (probably not necessary). Disable its cluster awareness with vgchange -c n and then re-enable it again, so it writes the bit again.
vgchange -c y
This should work , I think.
Greetz Jasja.

[RAC_1]

halt the cluster then start again. that's all

in case required do cmhaltcl

vgchange -a n your_vg
vgchange -c n your_vg
vgchange -c y your_vg
start cluster

[ASSIST]

Moreover you must to do

cmhaltcl
vgchange -a e
pvchange -t 0 /dev/dsk/
vgchange -a n
cmapplyconf -k -v -C
cmruncl >cluster>

Good luck!

[Ian Killer_1]

All excellent responses thanks. THe first response was sufficient and timely. Thanks to all for your assistance.

Cheers.