DL380 G4 - logical drive queued for rebuilding

[prxhi]

RAID1
2 Physical Disks
U320 146GB

Replaced one drive that was in "predictive failure" online. A day and a half later, the array still does not want to rebuild.

Restarted server and hit F1 to start data recovery. Boot to Windows 2003 Standard, log in to ACU and my array shows (Ready for Recovery) Logical Drive 1.

Do I need to force recovery of the array somehow? I changed the controller settings to "High Rebuild" priority but that didn't seem to do much.

I've attached a copy of the ADU.

Any help would be greatly appreciated.

Thanks,
Mike

[gregersenj]

Rebuild should start within a few seconds, and you can see it on the HDD - The on-line LED is flashing.

I couldn't open your ADU report.

Things that I know, may caurse the rebuild to start is:
Old FW versions on Smart Array Generation 6, may caurse a drive to be assigned an incorret SCSI ID, That might not excist, and the new drive get the correct ID.
The array controller will forever waite for a drive to be insertet on the incorrect address.
In this case you can try to fix it with a fw upgrade, wich isn't likely to fix the problem.

Another problem is "drive not be stamped for monitoring" You can google it, and find a possible solution.

I might remember incorrectly, but years ago, I remember, that using ACU to check rebuild progress, caursed rebuild to restart.
Try to check using the system management homepage. (I could be wrong on this, or it could be an old issue only.

BR
/jag

[prxhi]

Thanks gregersenj.

I've confirmed that the hard disk light on the new drive I used to replace the bad drive is not blinking.

I went back and checked the event logs and found the following.


Source: CPQCISSE
Event ID: 24603

Recovery of logical drive 1, configured on
Array Controller [Embedded], was aborted while rebuilding physical drive with DEVICE ID 0 on Port 2 due to an unrecoverable read error. The physical drive reporting the error is DEVICE ID 1 on Port 2.


Now port 2 contains the "good" drive that I want to rebuild off of. I've restarted the server and received the same error on 3 separate occasions.

So it appears that my "good" drive is now bad in that is reporting an unrecoverable read error whereby it is unable to successfully allow the spare drive to rebuild itself off of it.

Has anyone encountered this situation or received this error before?

I've also re-attached my ADU, zipped.

[gregersenj · Accepted solution]

I'm not a shark, on reading the ADU report.

But if I read the ADU report correctly.
You have replaced Drive 2:0, and rebuild was aborted due to read errors on the source disk 2:1.

Hopefully, some one else, will take the time, and check your ADU report, and confirm if I'm right or wrong.

If I'm right, then you have been struck by very bad luck, and got a multible drive failure within the same array.

[marcus1234]

,

some times firmware can cause this but you have firmware 2.84 which is extremely stable...

shutdown server remove cache battery on controller reboot

take option to continue its just saying it missed abit of data , that was jammed in the battery,,


proceed it it rebuilds now okay if not then a backup is required from tape..


only these options..

enjoy,

points appreciated..:)

[prxhi]

Just wanted to send an update in case anyone else was experiencing this problem.

It turns out that the remaining "good" drive in the mirror had bad sectors. These bad sectors were the reason for the read errors encountered during the rebuild process.

It would be nice if ACU employed a "best effort" option to skip failed sectors during an array rebuild. But then again that would result in an imperfect mirror copy which could be bad depending on the data that was stored there.

Fortunately, the sectors that were deemed bad did not contain any valuable data. We took a full image copy backup of the logical drive, plugged in 2 new drives, rebuilt the mirror, and restored the image to the new mirror.

I will also note that image software picked up on the bad sectors and we just skipped over them when taking the image backup. We also moved all the drives on this array controller to another server before going with our image backup method so that we could be sure and rule out any server related hardware.

[gregersenj]

Yes, IT's a bad situation.

But you were in a catastrofic HDD failure situation. Wich is when you have got a multible HDD failure, that will caurse data loss.

You can read aboute this in the Smart Array user guide (At least for the Smart Array 641)

An idea just come to my mind.
The degraded disk, is still oprational.
It might be possible to rebuild, from the degraded disk to a spare...maybe.

Thanks for the update, and sorry I have been away for a long time.

cheers
/jag

[prxhi]

thanks again for all your help, points assigned!

[gregersenj]

Merry Christmas.