topic Re: ¿Modify CLUSTER_LOCK_VG = down cluster? in Operating System - HP-UX

¿Modify CLUSTER_LOCK_VG = down cluster?

Jose Zamora — Thu, 05 Jun 2008 23:24:23 GMT

Hi:

I have in 2 computers ServiceGuard and HP-UX.
Modify CLUSTER_LOCK_VG it consequently has the fall of the cluster.

Thanks for its answer.

José

Re: ¿Modify CLUSTER_LOCK_VG = down cluster?

Jeeshan — Fri, 06 Jun 2008 03:46:19 GMT

if you modify cluster lock vg you need to rebuild the cluster.

Re: ¿Modify CLUSTER_LOCK_VG = down cluster?

Frank_W — Fri, 06 Jun 2008 06:45:10 GMT

Yes,

you have to edit the cluster-config file and check it with check it with "cmcheckconf".
This should work with a running cluster.

Then you have to halt the cluster and apply the new config with "cmapplyconf".

Re: ¿Modify CLUSTER_LOCK_VG = down cluster?

Asif Sharif — Fri, 06 Jun 2008 09:12:22 GMT

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=949610

Regards,
Asif Sharif

Re: ¿Modify CLUSTER_LOCK_VG = down cluster?

Avinash20 — Sat, 07 Jun 2008 02:33:31 GMT

Please refer to doc UXSGKBRC00017956 from www.itrc.hp.com
This is an excellent doc which will answer your queries.

Modify CLUSTER_LOCK_VG
This requires downtime, since we need to rebuild the cluster.

I assume that you have created the VG and have vgimported it the other node.

Locate the cluster configuration file. It is normally located in /etc/cmcluster on at least one node in the cluster. If unsure of the name of the file, try this command in
/etc/cmcluster:

$ grep LOCK *
CONF:# Lock Disk Parameters. Use the FIRST_CLUSTER_LOCK_VG and
CONF:# FIRST_CLUSTER_LOCK_PV parameters to define a lock disk.
CONF:# The FIRST_CLUSTER_LOCK_VG is the LVM volume group that
CONF:FIRST_CLUSTER_LOCK_VG /dev/vg05
CONF: FIRST_CLUSTER_LOCK_PV /dev/dsk/c2t5d6
CONF: FIRST_CLUSTER_LOCK_PV /dev/dsk/c2t5d6

If multiple files are reported, create a current configuration file:
# cmgetconf cluster_new.ascii

Edit your cluster configuration file with the name of the intended cluster
lock VG and disk

Edit your PV annd VG appropriately
Halt the cluster and update the cluster binary file:
# cmhaltcl -f
Decluster the VG
# vgchange -c n /dev/vg01
Activete the lock VG in one node
# vgchange -a y vg01

Update the cluster binary file, deactivate the lock VG and start the
cluster:
# cmapplyconf -f -C
# vgchange -a n vg01
# cmruncl