HI everyone,

I have IBM Flash Storage 5000 and it is running code 8.6.0.0

Unfortunately there is controller failure and we replaced with new one. But new node is in candidate and it is running code with 8.4.0.6.

We had removed the faulty node from system(CLI # rmnode) .

Kindly advise how can join candidate node to existing cluster of storage.

Please do needful

Thanks & Regards

HUJEFA

Hi,

usually the Storage Virtualize Code takes care to automatically join the node into the cluster, as long as the new node is in a clean candidate mode. You need to make sure that the new node does not show any open node error code (3-digit) that needs to be troubleshooted. You can doublecheck this in the service assistance gui.
If the code version of the new node is different to the cluster code, the new node needs to be populated with the codeversion that matches the clustercode. There are various ways to achieve this (e.g. manual codeupdate via service assistant gui, or automatic codeupdate during a manual addnode command etc.) - but before you proceed with this I would advice you to consult IBM support and upload a snap for review, so that IBM support can check for any obstacles and advice you on the best approach - especially since you seem to run 8.6.0.0 code that may be exposed to some unfixed issues that have been patched in a later codeversion.

E.g. be aware that there was a bug in 8.6.0.0 as far as I know that messed up the node rescue (another way of populating the new node with the cluster code) - see SVAPAR-100127 https://public.dhe.ibm.com/storage/san/sanvc/release_notes/86_releasenotes.html 

So my suggestion would be to raise a support ticket with IBM and let them review the snap data to get the best advice. 

------------------------------
Philipp Wachtler
------------------------------