Hi all,

 

I'm a long time UNRAID user, but a linux rookie, so I need your help.  I have a stable server running version 4.7 and I can no longer write to one disk.

 

General info:

Parity - 2 TB Hitachi

Data disks - 16 x 2TB (mix of Hitachi, WD, Seagate, & Samsung)

 

I recently upgraded 3 disks from 1 TB to 2 TB using the following process:

1)  Ran parity check - no errors

2)  Upgraded disk 2

3)  Ran parity check - no errors

4)  Upgraded disk 4

5)  Ran parity check - no errors

6)  Upgraded disk 10

7)  Ran parity check - no errors

This process completed last Sunday (12-2-12)

 

On Tuesday (12-4-12), I tried to rename a file on Disk 11 and received this error:

"Destination Folder Access Denied"

This was from a Windows Vista computer.  I receive the same error whenever I try to perform any write operation on Disk 11.  This is true from Vista, Win 7, Win XP, or MC via a telnet session.  The files on this drive are all movies and tv shows which seem to play normally.

 

Things that I have tried:

1)  Rebooted the server - no change

2)  Replaced Disk 11 with a new, precleared drive.  The rebuild and parity check seemed to proceed normally.  I can still access all of the data, but I still can't write to the drive. 

3)  Ran reiserfsck --check /dev/md11  - no change

4)  Rebooted again - no change

 

I have stopped adding anything to this server until I can get this fixed, because, I'm guessing that I'm not really parity protected if it can't write to Disk 11.

 

So, any help that you can give me will be greatly appreciated.

syslog_12-7-12.zip

The reiserfsck is failing with a segmentation violation. (it is attempting to use memory it does not own)

 

Step 1.  Perform a memory check on the server.  If memory is working properly, then it is a file-system fault that is causing the reiserfsck program to fail to complete. 

Step 2.  Try using

reiserfsck --fix-fixable /dev/md11

Thanks for the suggestions Joe.  Memtest has been running for about an hour with no errors.  I will let it run overnight and if it is still clean, I will try your second suggestion.

 

I ran Memtest over night (17+ hours) with no errors.  Then I tried reiserfsck --fix-fixable /dev/md11, then rebooted and still can not write to disk 11.  I have attached the results of reiserfsck and a syslog after reiserfsck.

Sorry, forgot to attach.

syslog_after_reiserfsck_--fix-fixable_12-8-12.txt

And here is a syslog after the reboot.

syslog_after_reboot_12-8-12.txt

Unless someone else, i.e. Joe, has another idea the drive will have to be removed. Try to copy all data off of the drive in preparation.

That is my next plan.  I seem to be able to read the disk fine, so I don't expect any trouble copying from it.

That is my next plan.  I seem to be able to read the disk fine, so I don't expect any trouble copying from it.

Once you have the data copied off, you can try running reiserfsck on the physical partition, although doing it that way does not preserve parity.

 

Eample: If the drive is /dev/sdg, you would run reiserfsck on /dev/sdg1  (note the trailing one on the device name, denoting the first partition)

 

Then, if it is able to complete, run a correcting parity check, expect errors, as parity is brought back into sync with the repair.

 

Joe L.

That is my next plan.  I seem to be able to read the disk fine, so I don't expect any trouble copying from it.

Once you have the data copied off, you can try running

reiserfsck --rebuild-tree /dev/md11

 

If that does not work, you can try it on the physical partition, although doing it that way does not preserve parity.

 

Eample: If the drive is /dev/sdg, you would run reiserfsck on /dev/sdg1  (note the trailing one on the device name, denoting the first partition.  The file-system is on the FIRST PARTITION... so make sure you run reiserfsck on the first partition, not the raw device for the entire drive.)

 

Then, if it is able to complete, run a correcting parity check, expect errors, as parity is brought back into sync with the repair.

 

Joe L.