4 node with diff patch level

[ASME]

Hi all,

i have 4 node MCG-CFS-RAC cluster all are with diff OS and Vritas patch level.

i am wondering how the things are running ? why they joined the cluster if the patch level are different?

how the patches are applied so that there is less down time ?

attaching the patch list o/p for all the nodes.

Reg,
asme

[sujit kumar singh · Accepted solution]

Hi,


nodes can join the cluster having the different level of patching for the OS As you also have said for Veritas .. all can be said that the difference might not have been that critical that would have allowed patches not join the clutser.


However , in your case i think you can see by taking a node out of the Cluster by properly stopping application on that. Apply the OS patches and the Veritas patches and make that join the cluster back and start the application on that.

if this is successful you can with caution do that for the Other nodes one by one.


Strongly suggested to have an Ignite backup of all the nodes before doing that and also please ensure that you are having always a reliable application databackup all the time.


Regards
Sujit

[Emil Velez]

They reason they are in the cluster is because it is serviceguard that is doing the clustering not the OS. In fact part of the guidelines for ServiceGuard is that it support rolling upgrades which means that if all nodes were 11.16 SG you could take a node out of the cluster, upgrade it to SG 11.18 add him back into the cluster and then do the same process to the other nodes.

Regarding veritas and OS patches the patches probably are not that critical but for stability you might want to make sure and get them up to the same patch level for stability especially with CFS.

[ASME]

Thanks guys,

how about using the "Dynamic Root Disk" instead of ignite ?

does DRD takes the image of patch level too ?

Regards,
asme