shutdown due to RAM upgrade

[S.S.]

Hi All,

Due to RAM upgrade to my HP 11i v1 machines it is required to shutdown the servers.

Actually my hp boxes are in cluster.

I would like to know before shutting down the server whether i would required to halt the package or simply i can do the following steps ?

as “root” user
# cd /
# shutdown –hy 0

Thanks all in advance.


Regards
Swetha

[Kapil Jha]

Anywhich ways shutdown script would run and ur package would be halted.
But its proper practice to halt the cluster and check if package is running on other node if you want that.
Then shutdown the machine.

BR,
Kapil+

[Johnson Punniyalingam]

>>I would like to know before shutting down the server whether i would required to halt the package<<<

You need to halt cluster before you proceed to shutdown the server for your RAM upgrade

use cmhaltcl -f

if you d'not halt the cluster and shutdown the server may think its panic reboot hence the Cluster services are up and running while you shutting down the server

Check below thread for more information

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1211644

[Horia Chirculescu]

Hello,

Stopping the cluster should be the safest way.

You can do it also with the cluster up, one machine a time (one node down then the other).

You must halt the node that you would need to upgrade, letting the package(s) to move to the other node (if it is the case). (cmhaltnode)

Afther that the node must be powered off in order to install aditional memory.

Afther upgrade, you can add the node back to the cluster using cmrunnode -v

Then repeat the procedure for the other node from the cluster.

Best regards,
Horia.

[Michal Kapalka (mikap)]

hi,

and what about the swap size ???

you should consider if you need to incerase the swap space.

mikap

[Taifur]

Hi ,

Follow below procedure
1. Firstly you have to halt package in u r actove node. (check man pages of cmhaltpackge)
2. shift package to ur passive node (cmrunpkg)

3.Halt active node.(cmhaltnode)
4.Shutdown server (shutdwon -hy now)
5.upgarde your RAM

Follow above procudure in ur passive node also

Rgds//
Taifur

[Gordon Sjodin]

You could see a memory dump if you do not stop the packages and cluster first on this node.

[Steven E. Protter]

Shalom,

You should not halt the package unless you want it to stop running.

If you do this upgrade one cluster node at a time, your applications should fail over with no loss of user service.

Schedule a downtime in case something goes wrong. An excellent chance to test your SG configuration.

SEP