Recently I've been having some trouble with one of my drives.  At first it was mounted as read-only and I ended up googleing solutions for that.  I ran reiserfsck -- check then reiserfsck -- rebuild-tree which I now know might not have been a good idea.  It still would not mount correctly and I was seeing errors in the log so I decided to remove it from the array, pre-clear the drive, and then re-add it to the array allowing the system to rebuild the data on the drive.

 

After re-adding the drive to the array it is now not mounting on boot-up and there are errors in the log.  It is also saying unformatted disks are present.  I decided it was time to post a question on this.  Any help is appreciated.

syslog-2013-12-29.txt

Which drive is giving problems? Post a screenshot so we can get a better idea of the current state of your array.

 

Disk 1 is the one giving me issues.

It looks like unRAID doesn't think there is any need to rebuild a drive, and disk1 is a new disk as far as it is concerned. Not sure how you got to this point, but I think the parity is valid for only the remaining drives, meaning there is no way to recover drive1. Did you let it recalculate parity without drive1? Can you provide more details about how you got to this point? Maybe someone else will be able to help.

 

Here is how I got to the point:

 

• Ran reiserfsck -- check
  
• Ran reiserfsck -- rebuild-tree as suggested in the prior step
  
• Stopped Array, Removed disk 1 from the array, Restarted Array
  
• Pre-Cleared disk 1 so it would show as new drive
  
• Stopped Array, Re-Added Disk 1, restarted array
  
• Disk 1 showed as a new disk, performed a data rebuild and expand as if I replaced a drive
  

 

That is pretty much what got me to where I am now.  With Disk 1 not mounting on boot but all the data seemingly available since I did have a valid parity drive before all of this happened.

All that sounds right, but I have never done reiserfsck. Can you still read the (emulated) drive1?

I can still read the emulated drive 1 contents.

I only did one drive rebuild (an upgrade) under 4.7 so I don't remember exactly how it went. You say you already rebuilt it but now it is showing unformatted? When you did the rebuild it didn't ask you to format then did it? I wouldn't think it would since the file system would get rebuilt anyway.

 

You may just have to wait on one of the more experienced users for further advice.

 

I only did one drive rebuild (an upgrade) under 4.7 so I don't remember exactly how it went. You say you already rebuilt it but now it is showing unformatted? When you did the rebuild it didn't ask you to format then did it? I wouldn't think it would since the file system would get rebuilt anyway.

 

You may just have to wait on one of the more experienced users for further advice.

 

Yeah when I did the rebuild and expand it prompted me to format the drive as well.  Which I did do.  Thanks for the help so far.  Hopefully others can assist as well.

See here: http://lime-technology.com/forum/index.php?topic=5072.msg47122#msg47122

I can still read the emulated drive 1 contents.

 

So if you go to \\Tower\disk1 in Windows Explorer you actually see your data?? 

 

That's very surprising, given the details shown on the Web GUI page.    If you ARE seeing data, it's almost certainly NOT "emulated" data ... but actual data on the drive.    You do NOT want to let UnRAID format it, as the data will then all disappear.

 

It seems that somehow you've got UnRAID to "think" that Drive 1 is cleared (i.e. it has a pre-clear signature); but it actually isn't.    So UnRAID just wants to format the drive, and then use it.    I suspect if you ran a parity check now, you'd get thousands (or millions) of sync errors, since that is not actually the case.

 

I would IMMEDIATELY copy all of your data from Disk1 to another drive on another system;  then let UnRAID format the drive; and then do a parity check -- which will probably have a LOT of errors, since the drive is not actually clear.  Then do another parity check (which should be error-free), and finally copy all of your data from Disk1 back to the system.

 

I was not exporting the disk shares so I turned that on and I can not see \\tower\disk1.

 

I removed the drive from the array once again and re-added it.  Now it is asking if I want to "Start - will bring the array on-line, start Data-Rebuild, and then expand the file system (if possible)."

 

Do I want to take that action to restart the array?

Yes, it now looks like it should with good parity and a missing Disk1 ... so doing the rebuild now will correctly rebuild Disk1.

 

It's off and running.  I'll update with the status once it completes.

 

It's off and running.  I'll update with the status once it completes.

If you did not see the data before, you'll rebuild the exact same lack of data again.

 

At this point, you might want to kiss that data goodbye.

 

It sounds as if you used the reiserfsck incorrectly, (not on the /dev/mdX device)

 

What disk device did you use it on?

 

Joe L.

There is a possibility the reason it is showing as unformulated is that the partition is pointing to the wrong start sector.

If so, let the re-construction complete (it should just be writing what is already there, complete with the file-system corruption that is preventing mounting)

 

Then, see this thread to perform analysis of what is on the disk, how it is actually partitioned, and what you can do to try to recover your data.

http://lime-technology.com/forum/index.php?topic=15385.0

3 of the drives are not configured AHCI.

Completed the rebuild and the drive was still not mounted.  I ran the following as suggested:

 

root@Tower:~# fdisk -lu /dev/sde

 

Disk /dev/sde: 1500.3 GB, 1500301910016 bytes
1 heads, 63 sectors/track, 46512336 cylinders, total 2930277168 sectors
Units = sectors of 1 * 512 = 512 bytes
Disk identifier: 0x00000000

   Device Boot      Start         End      Blocks   Id  System
/dev/sde1              64  2930277167  1465138552   83  Linux
Partition 1 does not end on cylinder boundary.
root@Tower:~#

 

Looking at the results I am guessing this drive isn't correctly partitioned but can anyone confirm?

3 of the drives are not configured AHCI.

 

I believe this is due to my motherboard treating 2 of the SATA ports as IDE ports.  One of my drives is actually an IDE drive so the fact that there are three makes sense to me.  Would it be better to remove those two drives from the on-board ports and put them on the AOC-SASLP-MV8 instead?

3 of the drives are not configured AHCI.

 

I believe this is due to my motherboard treating 2 of the SATA ports as IDE ports.  One of my drives is actually an IDE drive so the fact that there are three makes sense to me.  Would it be better to remove those two drives from the on-board ports and put them on the AOC-SASLP-MV8 instead?

 

Yes. It should not be an issue, unless it is.

It should not be an issue, unless it is.

 

:)

I was able to get the drive mounted.  I pre-cleared it again, formatted, and let it do a data rebuild.  Looks like I lost all the data on the drive (which I didn't have backed up  ) but it is back in the array.  Thanks for all the help.

I lost all the data on the drive (which I didn't have backed up 

 

http://lime-technology.com/forum/index.php?topic=31020.0