HALP! Corrupt Files, Directories -- CHKDSK ends with "unknown error"

[LittleNemoNES]

I have some files and directories I can't delete, move because they are corrupted.
I asked Windows to perform a full check on this storage Volume but it always ends with "unknown error" after having supposedly found and deleted some orphan files.

What are my options?

I have an ASS load of large files on this array and no extra space to move and rebuild.





::Side note::

I know everyone can have different experiences but I will not forgive Seagate for the 1.5 TB HDDs fiasco. Cursed be the day I decided to go with these drives for my array :-(
I've had 4 of them spaz-out or die on me with different controllers, PCs

 

[amdhunter]

Sounds like one of the drives might be failing. I hate to say this, but you should buy another drive and back everything up ASAP. That RAID setup will be a huge PITA to recover your data from if it decides to go down.

The only time I have run into an unknown error at the end of CHKDSK is when I have encountered a bad sector.

I have literally 50 or so drives next to my desk where I had performed data recovery on at work. Usually I'd give CHKDSK a go before trying anything else (which is a horrible idea BTW) just to be encountered by the unknown error message.

 

[LittleNemoNES]

Originally posted by: amdhunter
Sounds like one of the drives might be failing. I hate to say this, but you should buy another drive and back everything up ASAP. That RAID setup will be a huge PITA to recover your data from if it decides to go down.

The only time I have run into an unknown error at the end of CHKDSK is when I have encountered a bad sector.

I have literally 50 or so drives next to my desk where I had performed data recovery on at work. Usually I'd give CHKDSK a go before trying anything else (which is a horrible idea BTW) just to be encountered by the unknown error message.

Thanks for your say, amdhunter

some more color: The array became degraded because one of the 1.5 TB Drives dropped. I did some research online to burn the firmware ISO to disk. I rebooted without doing anything and that's when I noticed that my desktop was missing some recently placed files (my desktop is on storage drive).

I thought that was odd..

after yet another reboot (again nothing done to the PC) I saw chkdsk and it gave me the "unknown error" message @ the end.
That's when I decided to flash the last 2 drives that were w/ factory firmware to the latest (SDxB?). That went well and I began to rebuild the array -- it is healthy now and I can use it.

But the situation persists.
So the Disks seem to be fine but the file system is borked. What if I shrink the partition and create a new one? I definitely have more than 50% of the volume free (6TB total Space)

 

[Evadman]

Honestly, that sounds like a bad controller or something hosed in the OS not the disk itself. Did the controller throw any bad disk sector error messages? It may have marked a sector as offline uncorrectable, in which case that data is gone. But with RAID6, the same physical sector must have failed on 3 separate disks to get to that point. That is a several billion to 1 occurrence though. The controller (or perhaps controller and drive combo) failing.

Shrinking and growing the file system just moves the data physically on the disks, but it really sounds like your directory tree or file bitmap is borked. I would look for some other tools to recover it because my gut feeling is that the issue is with the OS, not the disk itself. Way too much has to fail to kill a RAID6 array, and the only major commonality is the controller (thus mainboard, memory,cpu too) and the OS.

If a shrink and grow works, then the issue is with the disk(s).

 

[taltamir]

AFAIK chkdsk does not play nice with raid, I heard a lot of people say "my raid array broke and then CHKDSK ran on my drives and it deleted all my data by "fixing" it"....

Undeletable directories and files are pretty common in NTFS, and chkdsk can NOT fix them, rarely some sort of tool like unlocker will work, but most often the only solution is to reformat. I just leave them until my next windows reformat (this is part of the reason you don't want NTFS as your fileserver's filesystem... really, what were you thinking?)

I recommend you get a real file system, like ZFS on open solaris or at least UFS on linux (or FreeNAS) for a fileserver.

Another advantage of ZFS is that it would have allowed you to find out if a drive is erroring, correct the errors without dataloss, and pinpoint WHICH drive is erroring. As it stands, all errors will cause silent data corruption and there is no way to tell which of your drives is crapping out (if that is even the case, it could just be random NTFS errors that have nothing to do with drive problems)

 

[LittleNemoNES]

Thanks for the advice, Taltamir

How can I reconcile this with me wanting to use a Windows OS?

Make a separate NAS server using ZFS?

If I use linux I should probably remove the areca controller, right?
Linux + HW RAID = no?

 

[taltamir]

i am not quite sure about HW RAID controller support, overall i prefer OS level raid (not "software" which is cheap card or mobo, but OS level) raid because it removes the single point of failure of the card. With ZFS raid6 (raidz2) i can reformat my OS Drive and reimport the array instantly with a single line command: #zpool import -f tank
I can even reformat and install a different os so long as it supports ZFS...

Making a seperate nas server is really the way to go, a cheap 64bit (must be 64bit for ZFS) CPU (I am using AMD X2 @ 2.4ghz) with 1 dedicated OS drive and then a raid5 or raid6 array (or if you have the money for it, raid1 or raid 10). If your dedicated controller is high quality hardware raid though, you can expect it to be slower.
You can also use different configurations than what i suggested, there is a lot of flexibility, i chose that one because its cheap and effective.