12H Auto Raid

[Nobody's Hero]

I have an HP 12H auto raid box with a failed drive. Fail over or redundancy protection worked fine. However, I am going to replace this drive that is bad. Can I just shutdown oracle so nothing is accessing the auto raid and then just swap the drive? Physically?

[Pete Randall · Accepted solution]

Robert,

You might want to shut down Oracle to be on the safe side, but the drives are supposed to be hot swappable. You've already said it's switched over to the other drive. You should be able to just pull the failed one and pop in its replacement.

Pete

[Helen French]

Hi Robert:

The drives are hot-swappable ! You don't need to shutdown anything. Just remove and install the new drive. The AUTORAID array will take care about everything else.

HTH,
Shiju

[A. Clay Stephenson]

Hi:

The main reason for having one of these things is so that you don't have to shutdown at all. Simply remove your bad drive module and replace it with the new module and walk away.
There will be a light show for up to 2 or 3 hours whiule the array re-balances but business can continue as usual during this period.

[S.K. Chan]

Replace the drive "hot", there is no need for shutdown. You can check the status of the rebuild using "arraydsp" command.

[Bill McNAMARA_1]

As the others are saying, you can remove the failed drive no problem... hot.

You can also place the replacement drive hot.

Configuration of the 12H that would affect the operation would be AutoRebuild functionality on or off.. and the rebuild priority may be high or low
AutoInclude may be on or off

Verify with arraydsp -a
where SN is the serial number returned from arraydsp -i

If AutoInclude is Off
The inserted disk will not be automatically included.
You will have to manually include via the disk changes option.

If AutoRebuild is off, you will have to rebuild manually.

If the rebuild priority is low, the rebuild will take longer, as the io to the array will get spindle priority over a rebuild operation.

You may prefer to do this during quieter period to avoid a higher hit on your application, in terms of performance, although, apart from that your application should not even know that a disk has failed or that data is rebuilding or migrating on the array.

If you plan to stop oracle, do a backup before rebuilding... just in case the rebuild fails... which it shouldn't !

Later,
Bill

[Sanjay_6]

Hi Robert,

With an autoraid, you don't have to shutdown the disk drive subsystem to replace the failed drive. you can do it online itself.

Check the disk array manual for more info on how to replace a failed drive.

http://h200004.www2.hp.com/bc/docs/support/SupportManual/lpg28364/lpg28364.pdf

Hope this helps.

Regds