ProLiant Servers (ML,DL,SL)
cancel
Showing results for 
Search instead for 
Did you mean: 

Array Ready for Rebuild Message

Peter J West
Frequent Advisor

Array Ready for Rebuild Message

Hi there,

We've got a Server with 2 x RAID5 volumes defined on it and when I view the status it says that one of the volumes is Ready for Rebuild.

I've rebooted the Server and hit F1 to have it attempt the rebuild but the message is still being displayed.

I suspect that Disk 0 in the volume has a problem based on the log file generated by the ADU.

Can someone please look at the attached file and confirm if this is the case for me?

If so then what options do I have? As I see it was have two problematic disks in the volume so we would need to backup the entire contents, remove both of the faulty disks, replace them and then recreate the array volume before restoring from backup. Is this correct or is there some other way around the problem?

Thanks for your help

Pete
4 REPLIES
MSECO_1
Trusted Contributor

Re: Array Ready for Rebuild Message

Hi Peter,

I just checked your log file.
Your Smart array 641 has 1.92 firmware revision, you need to upgrade to the latest one 2.74
Note: First you need to make shure there are no failures on your system before attempt the firmware upgrade.

http://h18007.www1.hp.com/support/files/storage/us/family/model/5559.html?lang=en&cc=us

For your hard drives, please download firmware maintenance CD ver 7.70 and boot from it, make shure that you have an UPS connected with your server
******************************************
Drive ID0 needs to be replaced (270 Hard read errors)

The rebuil was aborted on drive0 (02-07-2007)
The rebuil was aborted two times on drive0 (02-23-2007)

May I have your configuration:
There is only one RAID 5 with 6 hard drives and two logical drives or there are two RAID5 with 3 hard drives, each RAID with one logical drive without spare?
Logical drive0
Drives, slots?
Logical drive0
Drives, slots?

If you have an spare, were is it, slot# (0=left most)

Wich drive failed when this error shows?
Did you replaced any drives, slot#?
Keep learning!!!
Brian_Murdoch
Honored Contributor

Re: Array Ready for Rebuild Message

Hi Peter,

It looks from your report like SCSI ID0 won't rebuild because of read errors on SCSI ID2 and visa-versa.
In this type of situation the rebuild begins but fails after only a few % due to the read errors on an existing member of the raidset. This is where you get the "Ready for Rebuild" problem repeating itself. The controller can't reconstruct the new replacement drive due to hard read errors on another drive so it gives up.
In some cases a firmware update to disks and controller may help but it does usually result in a full backup and restore for peace of mind.

Get a backup as soon as possible and try firmware updates which might help. Once complete replace drive 0 as it has the most serious problems. Drive 2 might survive the rebuild of drive 0 but it has been generating timeouts and hard read errors (far less than drive 0 though).

Have a look at the analysed version of your report file (attached).

I hope this helps,

Brian




Peter J West
Frequent Advisor

Re: Array Ready for Rebuild Message

Hello there,

Many thanks for your suggestions and apologies for the delay in replying; i've been away on training for a few days.

Just to give a little more background the disk setup is like this:

Slot Disk Size Logical Unit
5 146GB Logical 2
4 146GB Logical 2
3 146GB Logical 2
2 146GB Logical 1
1 146GB Logical 1
0 146GB Logical 1

As you can see we have 2 RAID5 volumes and the fault lies on Logical Unit 1. Inspecting the arrays shows activity on all slots, apart from Slot 2 which is failing to illuminate. Slot 2 is the slot which has received a replacement disk but is failing to rebuild (due to errors on Slot 0 I believe).

As we now effectively have faults on 2 of the 3 disks in the RAID Set is it fair to say that the only way to recove is to backup the data, replace both disks, recreate the array configuration and then restore?

Unfortunately the faulty logical volume holds the OS, Exchange Server and all the mailboxes. So doing a rebuild as described above is going to be a somewhat time consuming process with no small amount of downtime.

Thanks for your assistance.

pete
Peter J West
Frequent Advisor

Re: Array Ready for Rebuild Message

Just a quick update.

I'm going to be dialling into work tonight and moving the Exchange Databases to the second array we have defined in the Server. At least that way if the main array does completely fail we will find it easier to recover.

I've also ordered 2 replacement disks for the faulty array and the plan is that at some point soon we'll schedule some downtime so we can do the following:

1. Backup the contents of C: and D: (both of which are on the faulty array).
2. Delete the array.
3. Replace the 2 x faulty disks (0 and 2).
4. Recreate the array.
5. Restore C: and D: (containing OS and Exchange Binaries).
6. Move the mail databases back.

A long winded and protracted route no doubt - but about the only way I can see to get a reliable and working system back in production.