Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

VMS Patching on 5 Node CLuster

 

VMS Patching on 5 Node CLuster

Hi Friends,

I am installing patches on a 5-node VMS 7.3-2 cluster, each node boots from its own disk. Just want to check on the best practises for installing patches for LAVC cluster members. I am planning to take one node offline from cluster and install the patches, test the node and put it back in the cluster. I will be installing the VMS_UPDATE_V15.00, PCSI v4.0 and other new ECO kits released in Mar'08, has anybody installed these kits, any problems faced ?

Thanks in advance.

Regds,
Sanjay

Cheers,
Sanjay Mundhra
I am a HPE Employee

Accept or Kudo

5 REPLIES 5
Hoff
Honored Contributor

Re: VMS Patching on 5 Node CLuster

"It depends."

For many cases, after reading the ECO kit release notes and here assuming that the cluster can continue to operate with one node down (at a time), the patch process is typically this:

perform a BACKUP /IMAGE of the system disk or otherwise ensure a path back out should the installation or the testing of the patch fail.

Ensure the prerequisites are installed.

Install the ECO.

Reboot.

Verify correct operations, isolating the node as required.

Repeat on the next node.

==

The BACKUP /IMAGE of the system disk is the path out if there are "any problems faced" that might be specific to your site.

==

In some environments, the above procedure and the above sequence is considered completely wrong and not best practices, and you cannot and should not test in the production cluster; you can and should test off-line, then roll the production cluster forward.
Jan van den Ende
Honored Contributor

Re: VMS Patching on 5 Node CLuster

Sanjay,

Hoff pretty much said it all.

Make REAL sure you have a GOOD system disk backup!
Since you are shutting down anyway: just make that a verified standalone backup.

... and you should really apply the patches to an independent test system, and test thoroughly, before doing the production environment.

And: _NO_ hurry. Do it at ease, in a quiet period. Take your time. Check and double-check. Preferable do it with TWO persons, and for EACH command, let the typed commands be checked by the other person before hitting . Prudence takes a cost, but it DOES pay off to avoid the trivial errors!

Success.

Proost.

Have one on me.

jpe

Don't rust yours pelled jacker to fine doll missed aches.
Duncan Morris
Honored Contributor

Re: VMS Patching on 5 Node CLuster

Sanjay,

be aware that there might be an issue with VMS732_SYS_v15 - first reported against 8.2/8.3 kits, but now Rick Dyson suggests that the same issue occurs in 7.3-2

see:
http://forums12.itrc.hp.com/service/forums/questionanswer.do?threadId=1218070


Duncan
Robert Gezelter
Honored Contributor

Re: VMS Patching on 5 Node CLuster

Sanjay,

In addition to the other comments, it is wise to remember that problems can occur differently on different nodes.

The mere fact that an upgrade has been successful on one node is no guarantee that the update will be without side effect on a different node.

- Bob Gezelter, http://www.rlgsc.com

Re: VMS Patching on 5 Node CLuster

Hi All,

Thanks very much for the responses. Customer has confirmed that the patches have been tested in their UAT (single node). I will be taking image backup on tape and adding 3rd volume shadow member and then breaking it for having a safe copy of current OS.

Cheers,
Sanjay
Cheers,
Sanjay Mundhra
I am a HPE Employee

Accept or Kudo