HP-UX 11i v3 upgrade

[Amrutha]

Hi All - Yesterday I upgraded the HP-UX OS from 11i v2 to v3. It sucessfully completed the OS installation. Now the system is up and running. I'm doing rolling upgrade and both systems in the cluster are running different versions. Can someone tell the steps to join the node into the cluster?

[VK2COT]

Hello,

a) Re-enable autostart on the upgraded node.
Edit /etc/rc.config.d/cmcluster:

AUTOSTART_CMCLD=1

b) Return packages running on the otehr node
back to this node and reset the AUTO_RUN and local package switching parameters (if
necessary).

Use cmhaltpkg, cmrunpkg and cmmodpkg...

Cheers,

VK2COT

[Raj D.]

Amrutha,

What SG version do you want to keep on the nodes.
Remember you have to do rolling upgrade, one at a time, keeping the package on other node and then fail back to avoid down time,

You said , you have upgraded 11iv2 to v3 then the SG also must be upgraded to 11.17 . And both the node should be in same SG level.

What is the output of:
# swlist -l product | grep -i servicegu

cheers.
Raj.

[Suraj K Sankari]

Hi,
As VK2COT said you check this parameter on this file /etc/rc.config.d/cmcluster:

AUTOSTART_CMCLD=1

if it 0 then make it 1 save the file.

As you upgrade the OS not SG so I don't think so there will be any problem.

you can start your cluster.

Suraj

[Duncan Edmonstone]

I find it odd that you would get half way through doing an upgrade without knowing how to complete it... still the advice given so far is fine, but remember that during a rolling upgrade of a Serviceguard cluster, all Serviceguard commands apart from cmrunnode and cmhaltnode must be run from the node with the later revision of Serviceguard installed.

HTH

Duncan

[Amrutha]

Hi All - Thanks for all your responses. When I upgrade the HP-UX OS the new devices has been created instead of using the existing device files. I didnt remove the fiber cables connected to the storage before os upgrade and it is not listed in the document too. Did it occured on your system or not?

[Rita C Workman]

If you went to 11.31 you likely have gone to SG ver 11.18. Depending on your version on 11.23, you may not be able to re-join the cluster. [But you don't mention what version your at...so...???]

If your new version is not compatible with the old version (I know 11.16 cluster will not allow a ver 11.18 node to join, not sure about 11.17).
So to get around this take the new ver 11.18 (0/S 11.31) node and create a NEW cluster. As you upgrade the old nodes, then simply add them to the new cluster until everything is moved over.

On that device numbers changed point you mention. That's not unusual. Remember the hardware path didn't change. So you should be just fine.

Rgrds,
Rita

[Amrutha]

I'm using HP storage that is NOT compatable. HP-UX 11i v3 requires active/active due to that I'm unable to do vgimport and created new device files.