The server has disconnected with an error. Server message reads: A protocol error occurred. Shell not supported on takenover partner

[#1]

I cannot ssh into my other filer. both are active/active.

the filer I can ssh into shows

(takeover)

for the other one when i ssh into it, it says

The server has disconnected with an error.  Server message reads: A protocol error occurred. Shell not supported on takenover partner 

any idea what happening?

[#2]

The cluster has done a takeover and the other node is down.  There is a virtual instance of the other node running in takeover mode so you can only ssh to the node that is running.  To get to the partner node you can toggle with "partner" or run commands "partner vol status" for example.  Do you have a case open?  You need to see why it failed over (core dump, hardware issue, etc.) then if ok to do a giveback perform the giveback...but work with support to confirm the partner can giveback and didn't have a failure.  If you have the SP/RLM/BMC setup you can get to the system console to see if it is waiting for giveback...or connect with a serial console on the controller to check.

[#3]

I believe the battery firmware was flashed last night. does it take a while before I can ssh back into it?

[#4]

You won't be able to ssh until after a giveback is done to that node... was the giveback done and failed or a missed step after the upgrade?

[#5]

NEVER do giveback without having console connection open, preferably to both controllers, but at least to the taken over controller. I have learned it the hard way …

For that matter never perform anything that requires controller reboot without having console. Fortunately, with modern hardware this is not the issue anymore (although sometimes it is hard to persuade customers to connect RLM) and does not require long trip to datacenter.

[#6]

Agreed. I wince and wait for at least a few "waiting for giveback" messages before a giveback. Even with the default check for giveback option on

Sent from my iPhone 4S