I discovered this evening that I have lost several hundred GB of data on my unRAID 4.7 machine. I'm hoping someone may be able to assist by either helping me recover the lost data, or confirming it's gone for good, so I can move on...

 

Here is the recent history leading up to the loss and discovery:

5/20 - One HDD (disk 3) is reported as "DISK_DSBL". No spare HDD is available.

5/26 or 5/27 - unRAID server loses power suddenly

5/28 - Data loss is seen via Failure Notification emails (but not discovered until today). Loss appears to have occurred between 23:47 & 00:47.

5/29 - New HDD is installed and preclear is started

6/1 - Preclear is complete and new HDD is added to array, array rebuild is started

6/1 - Array rebuild is complete, 600+GB discovered missing

 

My unRAID machine has three data drives, all WD 2TB Green. Lost data was approximately:

md1 - 47.9GB

md2 - 104GB

md3 (failed HDD) - 486GB

 

The original "disk 3" is still available, and has not been reformatted. It is not currently installed in the unRAID machine.

 

Please let me know what other data I can provide to help sort out this problem. Considering the loss appears to have been several days ago, after at least one reboot, would a syslog even be helpful? Looking at the syslog via unMenu, I rebooted early this morning.

 

Thanks for the help.

 

Dave

 

Well ... the FIRST thing you should do when a drive fails (as it did on 5/20) is turn off your server until you get a spare drive.    Otherwise, you're running "at risk" and if anything else goes wrong you'll lose data (as you've apparently done since then).    If you can't afford the downtime, then you should ALWAYS have a spare drive available for situations exactly like these.

 

Once you lost a 2nd drive, you no longer had good parity, so you could no longer do a rebuild of the failed drive.  It's not clear from your outline of events when (or even IF) you actually lost a 2nd drive.    When you got a new drive and installed it, did you try to simply Start the array?  If you had done that, and IF parity was still good, the drive would have simply been rebuilt and all your data recovered.    At this point, however, there's nothing you can do to recover the drive.    What you CAN do is attach the drive to another system; see if you can read any of the files; and, if so, copy them back to your UnRAID server.    If you attach it to a Windows PC, you'll need to install a Reiser file system reader ... this works well:  http://www.diskinternals.com/linux-reader/

 

What does the Web GUI show for the status of all of your current drives?

 

 

I did not experience a second drive failure, at least not that I was aware of. From what I understand about unRAID, that should have meant no loss of data, apart from what was on my "disk 3". I did have a spare drive, but it was called to duty elsewhere, unfortunately.

 

When I completed the preclear on the replacement drive, I added it to the array and started the array. The array was rebuilt. The data loss appears to have occurred prior to the rebuild.

 

The standard web GUI shows green balls for all drives, but I do see that my "disk1" has 126 errors listed.

 

I will try to mount the old disk and see if it is readable.

 

Thanks for the help!

 

Dave

I did not experience a second drive failure, at least not that I was aware of. From what I understand about unRAID, that should have meant no loss of data

 

That's correct.  If you only had one bad (red-balled) drive and good parity, then all you had to do was install a new drive, start the array, and UnRAID would have rebuilt the failed drive on the new one.

 

If, instead, you ADDED the new drive to the array ... assigning it to a different slot;  and then Started the array without rebuilding the failed drive, then you created a new configuration and your old data was destroyed in the process.  It sounds like that's what happened.  Fortunately, you still have the failed drive, so you can at least try to read it on another system and see if you can read at least some of the data.

 

all you had to do was install a new drive, start the array, and UnRAID would have rebuilt the failed drive on the new one.

 

I would then guess this explains the loss of data on disk3, the one that unRAID was not seeing. While I cannot exactly recall my actions, I know that the NEW third disk was not a member of the array, so I assigned it and then started the array. If I'm understanding correctly, what I should have done after the preclear was complete, was to just stop and start the array, and let unRAID take over. It is also possible I removed disk3 from the array prior to replacing. In hindsight, it sounds like a bad decision.

 

Does this also explain the loss of data on disk1 and disk2? This is the most confusing part to me. If the data was there, and then overnight is gone, I don't understand what happened.

 

I was able to mount disk3 on my desktop. Thanks for the tip on the DiskInternals utility. Looks like at the very least the disk is readable, so I can recover that data. unRAID was reporting it as unformatted, which I now know may not be the end of the world.

 

Again, thanks for all the help. Even if the data from disk1 & disk2 cannot be recovered, now at least I better understand the recovery process.

Dave

No, your actions don't explain the loss you saw on disk1  and disk2.  Not sure just what happened there, but it's probably not recoverable.  You'll just need to restore that from your backups.