How to remove a disk shelf.

Question

Greetings, I have a CX700 that is used for online archiving of old data. When I retired it from production duty and reconfigured all the raid groups, I kept the last shelf of disks unbound. These were going to be used as replacement disks. The problem I have is the system knew about the shelf. When I removed the link cables, the gui shows the array with a fault because a shelf is missing. If I plug it back into the loop, it still shows a fault because we have used a couple of the drives as spares and the disks are gone. I would like to get rid of the fault light on the front of the array and also the fault icon on the management gui. One of our CE's thought I should be able to unplug the shelf, and then reboot the array. It would come up without the disks and be happy. This didn't work and doesn't seem like it should anyway. How to I tell the controllers the shelf isn't there anymore and ignore or stop monitoring it? Thanks, Jeff

RobertDudley · Answer

Unplug the DAE connections and restart the mangement services; should clear the fault (power cycling the array should have cleared it as well).

AranH1 · Answer

Jeff,If there were no bound disks in the enclosure then removing the enclosure would be accomplished by restarting the management server as Robert pointed out. Are you sure there were no RAID Groups on that enclosure or hot spares?

jeffc2 · Answer

Things have cleared up now. It looks like it was a combination of factors. I did find there was a raid group with nothing in it bound but containing these disks. I deleted the raid group and restarted the management server on both SPs. It was still unhappy.

When drilling down to the icons I found everything in bus 3 enclosure 3 showed as empty except for the LCC cards which both showed as faulted. I verified there was no cable coming from bus 3 enclosure 2 and restarted the management server again. Still faulted. I finally powered off both cabinets (240 drives). Once it came up it seemed happy now. I'm guessing my first power cycle would have worked except for the disks being assigned to a lun but not having any of them bound. The raid group did not show any faults which is why I passed over it the first time.

Thanks for the help.

Jeff

AranH1 · Answer

Thanks for posting back Jeff. Good to know what the problem was. This is usually a painless process Aran

CLARiiON

Was this post helpful?