Unformatted disks

[tafkap]

Hi all,

 

i'm running under version 4.5 beta 7.

 

Today, i have 4 disks unformatted, i submit to you my syslog file, the main problem is : Tower emhttp: _shcmd: shcmd (37): exit status: 32

 

Thanks for your help.

 

[RobJ]

Situation does look very serious, you were right to stop immediately, although I don't think you should even have started the array with 3 unformatted disks.  I would not proceed further until we can all come up with the best course of action.  It is tempting to think you may have been seriously vandalized, as it is hard to think of a rational scenario that could possibly explain the situation you are in now.

 

What we need to start with, is a detailing of what has transpired recently, both what events and error messages you saw, and what actions you have taken (we know you upgraded to v4.5-beta7), what drives were installed or reconnected, and their order.  Do you have a backup of the config folder?

 

What I can see of the current situation:

* you are recovering from a crash, transactions were being replayed (Disk 4, 5, 6, 8 ), and a parity build began, which you stopped after a minute

* super.dat is corrupted, but it and disk.cfg were partly readable, so a drive assignment table is available

* 3 drives are seriously corrupted at the beginning, with no MBR and partition table, and no Reiser file system recognizable (Disk 2, 3, 7)

* 1 disk with seriously corrupted Reiser file system (Disk 1)

* 1 drive could not be found in its old slot position, you assigned it later to Disk 3, so that is what it probably used to be; the fact that it was missing is usually because drives were removed, failed, had cables reconnected differently, or you had upgraded to a different unRAID release with different Linux kernel - none of which makes sense to do at the same time as a serious crash (unless you like pain)!

 

Because of the corruption on Disk 1, it too may look unformatted, and it could not be mounted.  That makes 4 unreadable disks.  What is supposed to be on all of these disks?

[tafkap]

Hi Rob,

 

thanks for your answer.

 

Two days ago i had disk3 no detected, i supposed it was broken, so i bought 2 disks 1To for replace this disk and replace other disk whose size was small. So i replaced this disks and the result is before your eyes : a disaster. I started the array, stopped it, press the button Restore, but never formatted.

 

Today, my hope is not lost disk1, disk2 and disk4. If possible.

 

I tried to read the disc 2 on my PC with drivers to handle the format Reiserfs but my PC crash when i browse folders, i'm under Vista Pro. If possible i would like to do this method for disk1, disk2 and disk4 for save my datas.

 

Cheers

 

 

[tafkap]

Hi,

 

finally i could recover my data on disk2 with the software UFS Explorer, the disk1 is crashed, UFS Explorer can not read the files system.

 

If i restart my unRAID server since the beginning - i start with just two disks : one for the parity, the other for the data - can i recover disk1 from the parity disk ? if yes how can i proceed ?

 

And after, i reintegrate the others disks, and i launch a parity check.

 

Possible ?

 

Thanks

[tafkap]

Nobody for help me ? 

[limetech]

From the looks of things, it appears you do not have valid parity, so you probably will not be able to recover a completely failed drive.  However you may be able to restore the file systems.

 

What you should do is boot the server but don't Start the array.  Next, via console or telnet session, run 'reiserfsck' program on each data disk.  To do this you need to look on the Device page and find out which device identifier is assigned to each disk.  This will be the string inside parenthesis.  For example, you will see (sdd), and (sde), etc.  The command you want to use is this:

 

reiserfsck /dev/sde1

 

Notice that I took the device identifier from above and added the numeral '1' to it.  What this command does is check the reiserfs file system on partition 1 of the device.  You will notice after typing the command that it will prompt you to enter 'Yes' to continue.  You need to type it exactly like that - upper case 'Y' lower case 'es'.

 

Let it run - it make take several minutes, even up to an hour, for your larger drives depending on how many files are on the drive.  When complete, the program will either tell you all is well, or it will suggest you re-run with a particular 'switch', e.g., '--rebuild-sb'.  Follow the recommendations of the program.

 

Run this on all your data drives that should have data on them & then report back with the results.

 

How did you get into this state?  Power failure?

 

Edit: I should add - if you think Parity is valid, then follow directions here (similar):

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

[tafkap]

Thanks a lot, i'm going to do this.

 

Cheers

[lionelhutz]

Something sounds wrong in your description. You changed 2 disks at the same time? This should not be done. But still, that should not cause 2 other disks to fail.

 

Running the reiserfsck on those new disks is pointless. They are not valid disks yet.

 

It sounds like you have done a bunch more things besides just change 2 drives and attempt to start the array. I doubt you can recover disk 3 unless you can get the origional disk to work again.

 

Put all the disks back like they were before you began to mess with it. Unassign the parity disk. Try to get the drives to all work so you can assign them and start the array. Run the reiserfsck on the drives that have issues. Hope that the origional disk 3 problem was just a cabling or connection problem.

 

Let us know how you make out up to that point. Then, we can provide some possible suggestions for recovering the disk3 if it's still not working.

 

Peter

[tafkap]

Hi Lionel,

 

thanks for your answer.

 

Finally, I do not care about the disk3, it's just movies. I just want to recover disk1, for now.

 

What suggestions can you provide me for this point ?

[RobJ]

Disk 1 reported a seriously corrupted Reiser file system, so follow the instructions in Tom's post for it.  You should be able to recover all or most of the files on it.

[lionelhutz]

You'll have to run a check on it and hope it recovers.

 

You can't recover disk1 if disk 3 is gone. unRAID allows for a recovery from one failed disk.

 

I don't know if you can recover disk1 now. From your description it sounds like you have done some "Restore" button presses and likely have tried to rebuild parity and who knows what else. I can tell you that you have no chance of recovering disk1 from parity unless every other disk is in the array and working properly.

 

Your post of what you did suggested you replaced 2 data disks and started the array. Then, you stopped it and pressed the restore button but never formatted. Then, with no further description are we to assume you stopped and posted?  You have suggested you've removed the drives from the unRAID box and attempted to access them on a windows machine. The problem is that this description of events makes no sense. You can not change 2 data disks and then just "Start" the array. The array can only be started if one data disk is missing or changed.

 

Peter