Controller replacement on FAS2552

[IlirS]

Hello,

I have to replace a controller that has Ethernet ports down.

Is there any replacement guide or suggestions regarding controller replacement on FAS2552 dual-controller Clustered Mode ONTAP 9.1 ? It is an old system.

Is firmware update mandatory or not ?

BR,

Ilir

[SpindleNinja · Accepted solution]

I'd recommend engaging support if this controller is under support.  

here's the doc for the 25xx 

https://library.netapp.com/ecm/ecm_download_file/ECMP1147156 

[IlirS]

Yes, but the controller is not under support contract. It has expired and they are trying to renew it.

Thank you for your recommendations.

Ilir

[paul_stejskal]

9.1 is fine for that. I would recommend updating ONTAP when you get support entitlements though because 9.1 is at or will be EOVS. Also there are some security vulnerabilities that need fixing.

[IlirS]

Yes, I will as soon as the support contract will be renewed.

Regarding the "Replacing the controller module" guide, on page 46 it is suggested to perform the giveback process immediately when the replacement controller is fully seated:

"Enter one of the following commands from the healthy node’s console and wait for the giveback to complete

In Data ONTAP 8.2 or later: storage failover giveback – ofnode impaired_node_name".

Should I execute the giveback process at this moment when the replacement controller node has not passed the diag tests and other configurations ?

BR,

Ilir

[paul_stejskal]

I would check the SP or console port.

[IlirS]

Sorry Paul, I don't understand.

[paul_stejskal]

Sorry was too vague. Basically I would monitor the startup of the replaced node and make sure it gets to the "waiting for giveback" screen ok, or is ready for bootup.

It may  need the giveback attempt to try to initialize the node.

Support will be able to guide you through it if you are stuck when you get to that step. If you have enough entitlements, we can even send a field engineer as it is a hardware failure.

[IlirS]

I managed to replace the controller. I did the giveback procedure only at the end when all the diag tests and config restorations completed.

Everything went well except one NFS LIF which doesn’t work if I migrate it at its home port (new Controller B) and works only on partner’s port. NFS is used for mounting one datastore to a ESXi host.

When I migrated the LIF to Controller B the datastore became inactive.

Any recommendation for troubleshooting this issue ?

[SpindleNinja]

is the port healthy, up, cabled correctly?   Generally a lif will warn if you're trying to migrate it to a port that's not up/healthy.   

[IlirS]

Yes, the port is up and healthy. I can ping LIF IP Address when the LIF is migrated to its home port (Controller B) but the datastore is inactive. Maybe it has any old reference related with the faulty controller ?