1824977 Members
3881 Online
109678 Solutions
New Discussion юеВ

Re: Changing lock disk

 
Dave Forber
Occasional Advisor

Changing lock disk

We currently have a 2 node ServiceGuard cluster sharing the same external SAN disks, and using the lock disk strategy for control.

I'm going to be adding some new luns at the weekend and as part of this would like to change which disk is the lock disk.

Can anyone tell me the procedure for this?

Thanks.

4 REPLIES 4
Stephen Doud
Honored Contributor

Re: Changing lock disk

If you can open this technical document:
UMCSGKBRC00007864
TITLE: Changing the Cluster Lock Physical Disk
Dave Forber
Occasional Advisor

Re: Changing lock disk

For some reason, I can't open any of those links. As for the technical document, where would I find that please?
D Block 2
Respected Contributor

Re: Changing lock disk

Dave,
poke around in:
http://www.hp.com/go/ha
http://docs.hp.com <- at the Search: enter: lock disk

I pulled this list item(s) from Clay's first link above:


http://www2.itrc.hp.com/service/iv/docDisplay.do?admit=-1335382922+1120437650301+28353475&docId=prodITRC%2FDE_SW_UX_swrec_EN_01_E%2FServiceGuard.pdf



How to initialize the cluster lock disk(s) using cmapplyconf:
├в ┬в Halt the entire cluster.
# cmhaltcl ├в f
├в ┬в Perform the following command from all nodes in the cluster to remove the cluster
flag from cluster lock VG(s).
# vgchange ├в c n
├в ┬в Activate cluster lock VG(s) on one node only:
# vgchange ├в a y
├в ┬в Perform cmapplyconf on the node where you activated the cluster lock VG(s). The
cluster flag is added back to the VG automatically.
# cmapplyconf ├в C

├в ┬в Perform vgcfgbackup to backup the cluster lock information:
# vgcfgbackup
├в ┬в Deactivate cluster lock VG(s):
# vgchange ├в a n
├в ┬в Run vgcfgbackup on all other cluster nodes also:
# vgchange ├в a r
# vgcfgbackup
# vgchange ├в a n
├в ┬в Restart the cluster:
# cmruncl
internal).


final point: if you really really need to change the Lock Disks to a new Lock Disks? I'll assume you are using LVM..

on each node verify you are dealing with the exact same LVM PVRA. run this on each node for the shared-lock-disk:

echo "0x2008?4X" | adb /dev/dsk/cXtYdZ

verify that the output on each node is the same PVRA's.. then you know you have the exact same device. (I'll assume that the instance-numbers are different for each system's disk device for the lck-disk).

ok, the other point is the Zoning, right..
any new LUNs sh
Golf is a Good Walk Spoiled, Mark Twain.