Sign in with
Sign up | Sign in
Your question

Another RAID failure thread. 3 of 4 RAID drives OK but array is FAILED

Last response: in Storage
Share
September 5, 2008 1:39:54 PM

I am using an ASUS P5Q-Pro motherboard with the ICH10R chipset and 4 320GB hard drives in RAID 5. After an issue with a faulty partially connecting modular power supply connector which fed all the hard drives, I'm left with 3 OK drives but the array showing FAILED. All hard drives are physically working 100%.

This is what the Matrix storage manager said before the incident.

Array 001 (RAID 5), NORMAL
Port 5 Drive 4 NORMAL
Port 4 Drive 3 NORMAL
Port 3 Drive 2 NORMAL
Port 2 Drive 1 NORMAL
Non-RAID Drives
Port 0 Non-RAID (boot drive not part of array)

After:

Array 001 (RAID 5) FAILED
Port 5 Drive 4 NORMAL
Port 4 Drive 3 NORMAL
Port 3 Drive 2 NORMAL
Missing
Non-RAID Drives
Port 2 Drive 1 OFFLINE (listed as member of Array 001 but OFFLINE)
Port 0 Non-RAID (boot drive not part of array)

Considering 3 of 4 drives are listed NORMAL array members, the array should be DEGRADED not FAILED.

Following some internet forum advice for a similar situation I removed Drive 1 from the array in the hope of changing FAILED to DEGRADED to let me salvage data off it but instead I now get

Array 001 (RAID 5) FAILED
Port 5 Drive 4 NORMAL
Port 4 Drive 3 NORMAL
Port 3 Drive 2 NORMAL
Missing
Non-RAID Drives
Port 2 Non-RAID (not OFFLINE but not part of array)
Port 0 Non-RAID (boot drive not part of array)

Stubbornly, the array is still listed as FAILED.

Notes: There was no option to mark Drive 1 OK after it came up as OFFLINE. The entry Missing contains no information. As at September 14, the latest 1304 motherbord BIOS is installed and most recent Matrix drivers/utility is being used. Data should be OK as the drive was not being written to when the power supply issue struck. It would appear the RAID array's meta-data is corrupt in some way.

Is there a way to force Array 1 to DEGRADED status using the 3 of 4 array drives marked NORMAL? Is there any other way to access data on the array or rebuild it?
September 6, 2008 5:38:27 AM

It sounds like you lost power to the drives long enough for the controller to mark the array as failed. With most RAID arrays there is a timestamp on the drives you always want that timestamp to be the same on all drive in the array. If 1 drive has a different timestamp then that drive would be offline, or missing, or not part of the array or some other message depending on the controller. With RAID 5 you can have 1 drive go offline and not lose the array. If you have multiple drives go offline then you may or may not get the data back. (more on that in a sec) With RAID arrays you also have metadata that is stored on each drive,which is the information that tells the controller all about the array such as level and status, as well as the actual data (the stuff you want back). So right now it is possible that your metadata is telling your controller that the array is failed. If you lost power to all drives and your timestamp is the same on all drives or at least 3 of them then your data may still be there and easily recoverd. In this case you would need to re-create the array EXACTLY as it was before and DO NOT initialize the array. Not all controllers will let you do this but most will.

So if your issue was that all drives lost power and at the same time, the controller marked the array as failed and the timestams on the drives are actually the same then all you may need to restore is the metadata. You do this by recreating the array and not initializing. The initialization it what will wipe the actual data. The re-setting up of the array will only re-create the metadata. This is why its important to make sure you set up the array EXACTLY as it was before, so the metadata is as it was before.

If however you lost different drives at different times and there is different time stamps on the drives chances are that the data is not the same either. In this case you would have a much harder time recovering data. In fact even if it were just slightly off you may very well recover the array and try to boot windows and it will come up wanting to run check disk and if you let it run you will see a TON of orphaned filed and its not likely that windows would even boot and if it did you would lose alot of data.

Hope this helps you.

Also it goes without saying that you try this at your own risk.
!