Patch Upgrade

[Trng]

Hi,

we have two servers in cluster(hpux 11.23,MCSG11.17 running Sesap) .one package(DB) is running on primary and another package (CI) is running in secondary node.we are palnning to upgrade base quality pack to december 2008 on secondary node first.after observing it for 1-2 week we will install it on primary.is there any consistency issue will come between clusters?as this is in production we dont want to install both node at a time.requesting your suggestions .....

[Prashanth Waugh]

Hi Trng,
Halt Serviceguard on the node the patch is to be installed on.

You can just failover the pkg to another node

# cmhaltnode

2) Install this patch on that node.

# swinstall -s

3) Restart Serviceguard on that node.

# cmrunnode

failover the pkg again.

4) Patch needs to be installed on all nodes in the cluster.

You perform the same setps for other nodes in the Cluster.
===================================
I am also doing the paching activity now. I will be back within 1 hr with details steps.

Regards
Prashant

[Trng]

Hi,

i dont want to disturb both nodes at a time(same day),even customer not allowing that.If both nodes getting disturbed during patching activity ,it is very difficult to bring back .what is the best practise..

[Suraj K Sankari]

Hi,

That not a big issue you can go for one-by-one but its recomended that all the nodes are having the same patches, you should patch the other server soon.

These are the stpes need to do..

move the package into other node
halt the cluster
if possible take 2 copy of ignite
install patch
reboot your server
start the cluster
shift back the package

Suraj

[Trng]

Hi,

"you should patch the other server soon:......that is what i just to know whether the soon can be extend to one week or more ????

[Prashanth Waugh]

Hi,

You can extend patching date on adoptive node. It will not create any problem, But in cluster environment the patching level should be same so it will not create any problem in crtical condition. Its depends up to you.
-------------------------------------
also we are doing the annual patching in our cluster environment. so we are doing the patching on adoptives nodes by some time extents. we are going to keep each node under obseravtion for 2 to 4 days after patching.
Regards
Prashant

[Ganesan R]

Hi Trng,

There is nothing harm to extend the monitoring server to a week or more. That is the preferred method too.

Your first step should be move the package to primary node. The only concern is performance of primary node. If the primary node can manage the load of both the package you can monitor for a week or two.

After a week, switch back the package(CI) and monitor for some time if the package is running fine after patching.

If everything is fine, do the same thing for primary node.

[Prashanth Waugh]

Hi,

Before patching take the backup of vmunix and system file. It will be helpful.

Regards
Prashant

[Suraj K Sankari]

Hi,
>>i just to know whether the soon can be extend to one week or more ????

Yes, you can schedule after 1week or 2week but its recommend that ASAP you can patch your other node.

Suraj

[Trng]

Hi Ganesan,

My plan is as below

1--Move the CI package from secondary to primary and do patching on secondary
2-once secondary patching completed ,brging back the CI package from primary to secondary and primary will have DB package

3-observe for one week and do the same for primary ..


Is it ok or any risk involved due to patching inconsistency between nodes ?