Rebuilding of RAID 5 taking too long?

[mjpnuh_1]

Hello,

We've been having problems with are Proliant DL380 G5. We had a disk go down in are RAID so we replaced it with an identical drive. However we replaced this last monday and still have this status message in the array config util:

Background parity initialization is currently queued or in progress on logical drive 2 ( RAID 5 in array B). If background parity initialization is queued, it will start when I/O is performed on the drive. When background parity initialization completes, the performance of the logical drive will improve.

Now the disks are only 146gb each so I expected this to be rebuilt by now.

Anyone have any suggestions as to how i can rectify this?

Thanks in Advance
Mike

[wobbe]

If possible don't boot the o.s. rebuilds go quicker if no o.s. running.
You could try changing the rebuild priority using UCU.
Check with wbem if there are no errors on the harddisks. Perhaps the harddisk is trying to fix some error and this can reduce performance.

[mjpnuh_1]

Hello Wobbe,

Thanks for responding, we can't take this offline because it's are exchange server (which may also explain why its taking a very long time to rebuild) what is the UCU you mentioned? I cant find any errors in any logs the only thing mentioned is the message we are getting from my original post.

Any suggestions?

Cheers
Mike

[TTr]

USU is a typo for ACU.
Yes rebuilding will take very long time if the server is busy with i/o. The raid card gives more priority to the server i/o than to the i/o needed to rebuild the raid groups. Changing the rebuild priority will help but it will impact your server's i/o.

Do you have evidence that the rebuild has started? Or it has even finished and you are seeing a bogus error?
Check out this thread (rebuild may have finished) http://forums11.itrc.hp.com/service/forums/bizsupport/questionanswer.do?threadId=54724

And this one: http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1156170
(look at this reply: Nov 28, 2007 06:38:57 GMT) the rebuild will NOT start unless the disks have been idle for 15 seconds. If you server is busy and the disks are accessed all the time, the rebuild will not start. I think if you change the rebuild priority even temporarily might kick off the rebuild process.

[mjpnuh_1]

Hello TTr,

We've changed the priority and will hold on and see what happens there.

We've also spoken to support and they want us to update the firmware and reseat the Cache Module....not confident thats the issue but will do it anyway.

Not sure if its of relevance but all the drives connected to the Controller were showing solid blue lights this morning, i hope that means its doing something.

Any other ideas?

Thanks for your response
Mike

[TTr]

The blue lights turn on when ACU (or ORCA in BIOS) is accessing the drives. Exit the ACU to see if the lights go out.

Does the activity light blink at all on the replaced drive? It would indicate that the array is rebuilding.

> they want us to update the firmware and reseat the Cache Module..
The reason they mention the cache module is because if it is not funtioning the array may not be able to rebuild the drive. You can check the status of the cache module from ACU. If you decide to upgrade the firmware you will need downtime. In this case before you upgrade the firmware, allow the server to boot up and get into the ORCA and see if the rebuild will start there. Allow it to rebuild, it will be rather quick when the server is not running. Then decide if you go ahead with the fw upgrade.

[mjpnuh_1]

In response to your Question, all the lights in the array are flashing (in a way that suggests the drives are being accessed) We will be resetting the cache module tomorrow and will try to avoid the firmware upgrade if we can.

Apart from that we'll have to try some other things.

Thanks for your suggestions
Mike