ARECA 1220 FAILED DRIVE/Degraded Volume?!

The disk can't read this particular block, so it can't finish the repair.  (You probably have an XOR problem on the stripe as well) The ONLY thing you can do is get the block number, and write to that block. (The data is lost anyway, and with luck, you won't have a file there).   The write will force the controller to remap bad blocks and recalculate XOR parity.   The event log should give you specific info on offset


In future make sure you regularly run data consistency check/restores!  Had you have been doing this, then it is likely this never would have happened.

How do I get the block number? When I get it, how do I write to that black? I checked the event log this is all I get.



Time Device Event Type Elapse Time Errors
2009-09-28 07:51:25 Proxy Or Inband  HTTP Log In    
2009-09-26 06:05:35 ARC-1220-VOL#00  Complete Rebuild 045:31:39  
2009-09-26 03:22:51 IDE Channel 3 Reading Error    
2009-09-26 03:22:42 IDE Channel 3 Reading Error    
2009-09-26 03:22:34 IDE Channel 3 Reading Error    
2009-09-26 03:22:25 IDE Channel 3 Reading Error    
2009-09-25 23:52:40 IDE Channel 3 Reading Error    
2009-09-25 23:52:25 IDE Channel 3 Reading Error    
2009-09-25 18:24:02 IDE Channel 1 Device Failed    
2009-09-25 18:24:02 Raid Set # 00  RaidSet Degraded    
2009-09-25 18:24:02 ARC-1220-VOL#00  Volume Degraded    
2009-09-25 10:02:36 Proxy Or Inband  HTTP Log In    
2009-09-25 09:56:58 Proxy Or Inband  HTTP Log In    
2009-09-25 07:50:09 IDE Channel 3 Reading Error    
2009-09-25 07:50:00 IDE Channel 3 Reading Error    
2009-09-25 07:49:52 IDE Channel 3 Reading Error    
2009-09-25 07:49:39 IDE Channel 3 Reading Error    
2009-09-25 07:49:31 IDE Channel 3 Reading Error    
2009-09-24 08:46:13 Proxy Or Inband  HTTP Log In    
2009-09-24 08:44:44 Proxy Or Inband  HTTP Log In    
2009-09-24 08:32:11 IDE Channel 3 Reading Error    
2009-09-24 08:32:03 IDE Channel 3 Reading Error    
2009-09-24 08:31:55 IDE Channel 3 Reading Error    
2009-09-24 08:30:00 ARC-1220-VOL#00  Start Rebuilding    
2009-09-24 08:30:00 ARC-1220-VOL#00  Abort Rebuilding 000:02:32  
2009-09-24 08:29:59 Raid Set # 00  Rebuild RaidSet    
2009-09-24 08:29:59 IDE Channel 1 Device Inserted  

Darn, I was hoping the event log gave you the block number.  Once you had it, then you could use a binary editor, or just write a simple program to write to it.  I have access to the Areca API as I am a RAID architect, and could write a program to force repair, but I would have to charge you.  You should be able to get a utility from Areca that gives you the exact block number.   So my advice is to contact them directly.   It sucks when a rebuild dies due to a bad block, and there is no obvious way to force a write and move on..  The data is lost anyway.

Based on results, you have numerous read errors.  So you could have XOR errors, hard read errors, or combination of both.   So recovery has to be meticulously done and require somebody who knows what they are doing.   Since the problem is happening in the middle of a RAID6 rebuild, then you can forget using one of those off-the-shelf RAID reconstruction software packages.  The RAID6 rebuild uses a proprietary layout and algorithms that mark what has and hasn't been done, and you will really muck things up if you even try to go down that path.

Best to contact Areca and hope they have a solution for you.  The alternatives are all expensive unless you just want to either recover from backup, or ghost the system as it is to a backup, rebuild the array from scratch, then restore.

(My opinion is that you will ultimately have to get some scratch disks, do a ghost backup, rebuild the array - with full data destructive initialization, then restore.   No matter what you do, do it quickly ... lose one more disk in the interim, and you have 100% data loss)

Since the error comes from drive channel 3, could I take that drive out, and also channel 1(which is the one it said that failed) and rebuild both at once? It says we can lose up to 2.