drive enclosure replacement

[trireed]

I have an 8400 with 12 drive cages, is it possible to replace a drive enclosure online without shutting the array completely down ?

I have moved all PD's data that reside on the cage but concerned if cage is removed before new cage is inserted will the other cages still be able to process data.

[MammaGutt]

I have an 8400 with 12 drive cages, is it possible to replace a drive enclosure online without shutting the array completely down ?

I have moved all PD's data that reside on the cage but concerned if cage is removed before new cage is inserted will the other cages still be able to process data.

If the system is correctly configured and healthy, all cages are connected thru two paths/loops which are reversed. An unrealistic example would be:

Loop1: node -> cage 1 -> cage 2 -> cage 3 -> cage 4
Loop2: node -> cage 4 -> cage 3 -> cage 2 -> cage 1

Lets say you remove cage 2, cage 1 is still connected thru Loop1 and cage 3 and 4 thru Loop2.

So replacing an empty cage (keep in mind that an empty cage is unsupported) is no problem. With you have HA cage you can also replace a cage without removing the disks first.

[cali]

Hi,
we have removed two Times a Drive Cage (and Dives) online.
After unalloc and dismiss PDs we use:

showcage
servicecage remove -f cage
showcage
checkhealth

Cali

[trireed]

Cali, so we replaced cage and all the original drives are now showing failed in the new cage, in my case cage12, they were in cage11 which was removed.

all drives were unalloc and also servicemag start was ran on each PD to move data off prior to removing cage11. All drives show normal on cage, firmware is current on cage. If you do a showpd though all drives are failed state.

Wondering if we are missing something here, we tried servicemag resume on each drive and all completed successfully but drives still showing failed state which is odd since it seems to be using drives. Looking for a way to reset drive status out of the failed state?

[trireed]

Found my answer !

If a drive has a failed status but you know drive is good, like what happened to me due to a cage replacement you can clear the error by issuing:

controlpd clearerr wwn

This is not documented in CLI reference since its a hidden command for support.