Arch Linux

CubeGod

Member

Registered: 2013-07-09

Posts: 120

Massive data corruption for (seemingly) no reason at all.

Okay, my suspicion is the disk being damaged (smart reports it as healthy, running a scan on it now, already did a complete binary dump on it to make sure I don't damage more data while recovering what can be recovered)

Anyway, some data on the drive, it's a bit over 4 years old, a Toshiba MK2576GSXN (laptop drive), it has 2.1 years of power on time.
As for the events leading up to the massive damage loss:
Do a system update (arch), reboot, everything is working fine.
check emails, browse the web a bit, reboot
select windows 7 at the grub prompt
play dragon age for a few hours, reboot.
back in arch, checking emails again, shutting down to go to bed.
when I get back I'm prompted with this:
http://i.imgur.com/LfLqVQl.jpg
Okay, that's odd so I try again thinking it's something random
http://i.imgur.com/KJhsVHY.jpg
Nope, but three's a charm, eh? (now I'm just hoping to get some more useful error messages since the two last boots had different output)
http://i.imgur.com/YP6MkLH.jpg
And again, hoping to get some sort of useful input, anything.
http://i.imgur.com/pe35al3.jpg
http://i.imgur.com/aN3IjVS.jpg

Oh and windblows was similarly broken claiming that the install was corrupted and it needed reinstall (asking for disk)
Anything besides hdd crash that could cause this?
Windblows didn't have access to the internet and the only third party software installed was Dragon Age: Origins, Fallout:NV and TES IV: Oblivion

moderator edit: The images are too large. Please use thumbnails or links. See Forum Etiquette: Pasting Pictures and Code and screenshot posting rules. Thanks. --fsckd

Last edited by fsckd (2014-06-24 15:20:01)

ooo

Member

Registered: 2013-04-10

Posts: 1,638

Re: Massive data corruption for (seemingly) no reason at all.

can you mount the broken partitions if you boot from arch iso, if not post the error you're getting.
It wouldn't hurt to know the filesystems in use. did you try fsck.<fs_type> already?

CubeGod

Member

Registered: 2013-07-09

Posts: 120

Re: Massive data corruption for (seemingly) no reason at all.

Better yet, I installed another drive, put arch on it and I'm right now restoring data from the binary dump (mind you some files are destroyed, most things I cared about are fine though)
the partitions themselves mount but they don't boot.

fsck didn't return any errors at all, which perplexes me (disk doesn't boot either way, still gives those errors, even reinstalled grub from archiso on it and it still has the same result)
As for files, some fail CRC checks, others just aren't even there. (personally just interested in what could've caused such corruption if it isn't hdd failure)
I personally don't trust this hdd further and will only use it for experiments.

alphaniner

Member

From: Ancapistan

Registered: 2010-07-12

Posts: 2,810

Re: Massive data corruption for (seemingly) no reason at all.

Did you do a long SMART test or just the quick test? It's very common for a drive to pass the quick test but fail the long test.

If it passed the long test, put it through the wringer by running a badblocks write-read test. The write-read test will destroy data on the target drive with no chance of recovery! You have been warned.

badblocks -sw /dev/sdX

That will probably take numerous hours. It was suggested elsewhere to configure it to keep running until ten consecutive passes have completed without error. This will probably take several days, and could potentially run indefinitely if it keeps finding errors.

You might also consider checking the warranty. I know WD and Seagate have some drives with a 5 year warranty, but I'm not so familiar with Toshiba.

---

But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner

graysky

Wiki Maintainer

From: :wq

Registered: 2008-12-01

Posts: 10,650

Website

Re: Massive data corruption for (seemingly) no reason at all.

@alphaniner - I have called badblocks through e2fsck I think like this for a non-destructive which did flag some bad sectors.  I then boxed out a few gigs before and after them and (/knock on wood) haven't seem problems using the same disk.

# e2fsck -cckv /dev/sdXN

---

CPU-optimized Linux-ck packages @ Repo-ck  • AUR packages • Zsh and other configs

alphaniner

Member

From: Ancapistan

Registered: 2010-07-12

Posts: 2,810

Re: Massive data corruption for (seemingly) no reason at all.

Thanks for the info. I did something similar years ago and never had any further issues. I think it's slower and less thorough than the [destructive] write-read test though, and OP doesn't seem to be concerned with trying to recover anything else from his drive.

---

But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner

CubeGod

Member

Registered: 2013-07-09

Posts: 120

Re: Massive data corruption for (seemingly) no reason at all.

Did you do a long SMART test or just the quick test? It's very common for a drive to pass the quick test but fail the long test.

If it passed the long test, put it through the wringer by running a badblocks write-read test. The write-read test will destroy data on the target drive with no chance of recovery! You have been warned.

badblocks -sw /dev/sdX

That will probably take numerous hours. It was suggested elsewhere to configure it to keep running until ten consecutive passes have completed without error. This will probably take several days, and could potentially run indefinitely if it keeps finding errors.

You might also consider checking the warranty. I know WD and Seagate have some drives with a 5 year warranty, but I'm not so familiar with Toshiba.

I've already run the long test, it passed.
as for badblocks destroying everything, in case I forgot any data I wanted I already have a 1:1 binary copy of the drive in its current state so that's fine (although I already pulled everything)
afaict the laptop had a 2 year limited warranty, the drive is thus no longer covered (either way I just replaced it with a WD blue anyway, I like WD. never had a WD drive fail, seagate and samsung are also good, toshiba drives are subpar in comparison)

I'll now proceed with nuking all data with badblocks, hopefully this will provide information.
If that fails I'm going to have to assume that bad firmware caused it (probably for the drive controller if I had to guess)

CubeGod

Member

Registered: 2013-07-09

Posts: 120

Re: Massive data corruption for (seemingly) no reason at all.

UPDATE: badblocks is still running o.o;
I am amazed that it's working for this long, not even a complete wipe of a disk (same size, same rpm) took this long. (This is not a bad thing, on the contrary I am positively surprised)

dice

Member

From: Germany

Registered: 2014-02-10

Posts: 413

Re: Massive data corruption for (seemingly) no reason at all.

As you mention Fallout:NV, I had a similar issue some days ago. I played Fallout for several hours booted back into arch did some email and stuff and shutdown.
Some hours later I booted Windows which told it did some filesystem repairs. After that Fallout said some dll was missing. Reboot. Booting arch.."kernel missing" or something like that..
booting windows..infinite loop of checking file system->reboot->checking filesystem->reboot...

For some reason Windows "repaired" my file systems to death...Thankfully I had a recent disk image lying around. After copying that to the disk everything works fine again.

So maybe some windows repair stuff happened in your case too?

---

I put at button on it. Yes. I wish to press it, but I'm not sure what will happen if I do.  (Gune | Titan A.E.)

alphaniner

Member

From: Ancapistan

Registered: 2010-07-12

Posts: 2,810

Re: Massive data corruption for (seemingly) no reason at all.

(either way I just replaced it with a WD blue anyway, I like WD. never had a WD drive fail, seagate and samsung are also good, toshiba drives are subpar in comparison)

I'm a WD fanboy myself. But I was troubleshooting a machine at work this week and two consecutive bad WD Blues had me tearing my hair out. And they were replacements for another bad WD Blue. They may have all been from the same batch though.

And since we're talking about FONV, I've also been playing recently and haven't had any fs issues. But I'm running XP and it's a single boot machine, and the game is installed on a network drive.

---

But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner

CubeGod

Member

Registered: 2013-07-09

Posts: 120

Re: Massive data corruption for (seemingly) no reason at all.

Okay, badblocks finally finished, literally took over 24 hours (is this even normal?)
anyway each one only had the output of done so not copying the output since it's pointless.

As you mention Fallout:NV, I had a similar issue some days ago. I played Fallout for several hours booted back into arch did some email and stuff and shutdown.
Some hours later I booted Windows which told it did some filesystem repairs. After that Fallout said some dll was missing. Reboot. Booting arch.."kernel missing" or something like that..
booting windows..infinite loop of checking file system->reboot->checking filesystem->reboot...

For some reason Windows "repaired" my file systems to death...Thankfully I had a recent disk image lying around. After copying that to the disk everything works fine again.

So maybe some windows repair stuff happened in your case too?

Well, I never had that issue with fallout... I was playing DAO.
as for a windows "repair" causing it, I find it unlikely, the first thing I do is boot into arch, when I shut down my system after the final reboot it was from arch and the first thing I did was to try and boot arch.

as for the message windows gave it didn't even give me a repair option, it asked for a disc though

I'm a WD fanboy myself. But I was troubleshooting a machine at work this week and two consecutive bad WD Blues had me tearing my hair out. And they were replacements for another bad WD Blue. They may have all been from the same batch though.
And since we're talking about FONV, I've also been playing recently and haven't had any fs issues. But I'm running XP and it's a single boot machine, and the game is installed on a network drive.

Oh? I've never had a WD fail but as with any mechanical device I know it's possible, however WD seems to be the best constructed ones (especially WD blacks, if I can afford it I always go for WD black)

Anyway, the drive comes out clean but my system goes ape whenever the drive is in the sata slot, could corrupt firmware cause this? (trying again after I wrote a complete binary dump back to it (including mbr) it does the same but as long as the hdd is in I can't boot usbs either, they all cause similar explosive behavior)
Either way this drive is dead in my eyes, will use it for experiments though

Last edited by CubeGod (2014-06-25 22:28:59)

kahrkunne

Member

Registered: 2013-08-24

Posts: 88

Re: Massive data corruption for (seemingly) no reason at all.

I can't think of anything outside a harddrive failure, but I'm not that experienced.
On a different note, people, please stop shilling for your favourite hard drive vendors. If anything, take this to the off topic section.