So I think I've made a few mistakes with my server. 

 

I physically cleaned it out and tried to add a hard drive. when I did this I noticed one of my old hard drives wasn't being noticed - the light on my Sata card wasn't lighting up unless I played with the cable so I figured bad cable. I replaced the SATA cable with another from my SATA card and also disconnected the new drive (this was not formatted or added into the array).

 

The hard drive that wasn't being noticed now came up with a red ball icon.  At this point I figured (wrongly I guess) that it was just a glitch so I unselected the device for it, rebooted, re-added the device and did the trust my array procedure.

 

The Red ball icon now came up blue and I formatted it. 

 

Unraid is now doing a parity check and has over 300 million sync errors. I guess I screwed up and it's now in the process of wiping my Parity drive, whereas I expected it to just rebuild the hard drive.

 

Is this fixable?

 

 

Post a syslog for one of the experts to look at and analyze.  For instructions see the first post in this thread:

 

  http://lime-technology.com/forum/index.php?topic=9880.0

I've attached the logfile.

logfile.txt

The process chosen told unRAID to not rebuild the drive, and then telling unRaID to format the drive was an instruction to clear it of data!  UnRAID shows a drive as 'unformatted' if for any reason it cannot mount it - often the drive is not really unformatted but instead has file system corruption that can be recovered using the reiserfsck tool.

 

I would think that the best chance of recovering any data from the problem drive at this point is to:

• stop the array and then restart it in maintenance mode
  
• from a telnet/console session run a command of the format
  reiserfsck --rebuild-tree --scan-whole-partition /dev/md?
  where ? corresponds to the disk number of the problem disk.
  

This command will read every sector on the disk trying to locate what looks like directory information.  When it finds any it will try and re-instate the file it belongs to.  If you are lucky then most of your files may be found.  You may also find a lot of files placed in the lost+found folder with cryptic names where the name could not be found.

 

However I am a bit concerned that when you tried to do the parity sync you got the number of errors you said in the opening post (the syslog only shows a small number).  I would have expected only a small number to occur

 

You may find it easier to recover your data from your backups (assuming you have them) rather than follow the procedure above.

Yeah I figured I messed up twice, Once on the trust my array and then on the format. 

 

There still might be a loose cable or something that is contributing to the errors and I haven't written to the drive so maybe I can recover the data. I will try it, I have nothing to lose.

 

I use highwater and user shares so the problem is that I just don't know what data I've lost, I did see that some TV shows have episodes missing so that is going to be annoying.

 

 

The process chosen told unRAID to not rebuild the drive, and then telling unRaID to format the drive was an instruction to clear it of data!  UnRAID shows a drive as 'unformatted' if for any reason it cannot mount it - often the drive is not really unformatted but instead has file system corruption that can be recovered using the reiserfsck tool.

 

I would think that the best chance of recovering any data from the problem drive at this point is to:

• stop the array and then restart it in maintenance mode
  
• from a telnet/console session run a command of the format
  reiserfsck --rebuild-tree --scan-whole-partition /dev/md?
  where ? corresponds to the disk number of the problem disk.
  

This command will read every sector on the disk trying to locate what looks like directory information.  When it finds any it will try and re-instate the file it belongs to.  If you are lucky then most of your files may be found.  You may also find a lot of files placed in the lost+found folder with cryptic names where the name could not be found.

 

However I am a bit concerned that when you tried to do the parity sync you got the number of errors you said in the opening post (the syslog only shows a small number).  I would have expected only a small number to occur

 

You may find it easier to recover your data from your backups (assuming you have them) rather than follow the procedure above.

 

I use v4.7 at the moment so no maintenance mode, any other way to do this?

 

Ok I found this for running Reiserfsck on 4.7

 

http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

 

I am now running reiserfsck --rebuild-tree --scan-whole-partition /dev/md?  Will update on result

 

 

OK - I think all the data is gone, Anyone have any tips for finding out what went missing, else I have to manually go through all the movies and TV and see whats not in the folder.

Ok I found this for running Reiserfsck on 4.7

 

http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

 

I am now running reiserfsck --rebuild-tree --scan-whole-partition /dev/md?  Will update on result

If there is no maintenance mode (I have no experience of 4.7) then you have to run it against the /dev/sdX1 where sdX corresponds to the device allocated (which you can get from the GUI).    Do not forget the 1 on the end to specify the partition.

 

Note that after running this that parity will be completely messed up so will need recalculating as if it were a new system.

As I was going to sleep last night, I remembered that you had added a new drive.  I made a mental note to check to see if it was a WD drive but the screen shot had been edited out of your original post.  If it was a WD drive you should probably read this: 

 

  http://lime-technology.com/forum/index.php?topic=36065.0

 

An inspection of your cables will tell if this might be an issue.