Storage HDD went straight into messy chkdsk

[BadOmen]

Greetings. After installing a new SSD and reorganizing the use of my SATA ports, everything seemed fine for two days. Then, on boot, chkdsk started out of nowhere, but apparently testing my storage HDD.
It fixed index problems, and now it's going file by file with the messages:
- copying orphaned file ____.__ from directory xxxx to xxxx (or something like that)
- insufficient disk space to recover lost data

This has going on for 16 hours and counting.

I had moved this drive from a Marvell controller to the native port, and it seemed ok.

Any tips? Should I stop the chkdsk and try to boot again?
It may be something obvious, but tension is not allowing me to think clearly.
Any input is appreciated. Thanks.

 

[LoveMachine]

I had a very similar situation a few weeks ago when installing a new SSD. I switched the HDD to a different port and CHKDSK did exactly what yours is doing. I have no idea what changed or why the errors suddenly appeared, but to put your mind at ease, once it finished, everything was fine and dandy with no file corruption that I've found yet.

 

[Elixer]

Once chkdsk has started, you have a very good chance of doing more damage if you reboot while it is doing writes.

While, normally, I would just leave it alone, and hope for the best, in this case, it has been going on for 16+ hours ?
Something is very, very wrong. Could be the controller on the external has gone bad, or the HD itself.
At this point, I would stop it, knowing that it will corrupt whatever it is trying to write, but since you are getting message about 'insufficient disk space' it really sounds like you don't have much of any choice.
I assume this is the 3TB HD ? Hope you had backups.

 

[BadOmen]

The funny thing is that I've never seen chkdsk being triggered automatically on boot to check on a storage drive without any OS in it. Makes me think if the SSD doesn't have any participation in this...

I have no idea what changed or why the errors suddenly appeared, but to put your mind at ease, once it finished, everything was fine and dandy with no file corruption that I've found yet.

Those are great news (and I needed them), I really hope it's the exact same situation you had. It's been some scary 24 hours. Did you wait until the end or rebooted in the middle?

it has been going on for 16+ hours ?

Yes, almost 24 now.

I assume this is the 3TB HD ?

Yep, 3TB but not external. It's been a very good internal storage since last year, when I bought it. Never gave me problems, until the SSD came and I moved the 3tb from native sata to Marvell, then I noticed that Marvell ports delay the boot process when used and took it back to another native port.

 

[LoveMachine]

My event was also after moving from Marvell to Intel port. I had about 2TB of data on mine (3TB drive), and it did take a while, maybe an hour or so, but not 16 hours. Hope everything turns out OK. It's been a few hours since your first post. How goes it?

 

[BadOmen]

When I got home last night, it was on loop with the message:
"Error in writing output log".
That stayed until the morning, when I finally hard reset the computer.
The aftermath:
- The affected partition is totally full, but I can't see what's filling it. Hidden chk files?
- my photos folder was wiped out, just 4 pictures were spared out of 500+.
- A subfolder of my mp3 folder was also gone.
- When booting from the SDD, a double-click in one home video opens another totally different video. Messy index?

When booting from my pre-ssd HDD, though, this last problem doesn't happen. What you click is what you get.

I have a backup of most pictures, and most likely have a backup of the mp3. I'm just afraid of choosing the next step now:
- Try some tool to recover the wiped files?
- Ccleaner to remove whatever is filling my drive?
- Seatools to see if the drive is really ok?
- Checkdisk again????

 

[Elixer]

If you want to try to salvage things, the best thing to do is clone the drive to another drive, and work on that one.
Test disk (free) has worked well for me in the past, and there are other free utilities that can try and get back your pics on the same site.

I would force explorer to show all files, and not hide anything, then check what the files are.

Once done with that, your best bet would be to format the drive. I would also check with crystaldiskinfo to make sure the drive is still reporting good values for SMART.

 

[Russwinters]

What kind of 3TB? WD, Seagate?

(works in data recovery field as an engineer)


We have been seeing tooooons of 3TB they are dropping like flies.


CHKDSK running on a non boot drive is bad news, stopping it midway wont make it better, but it wont make it much worse, either, TBH.

(edit: CHKDSK running on ANY drive is bad news, if there is important data on a drive, NEVER run chkdsk if it wants to, copy your data out using 3rd party utility, etc before you run it.)

IF its been running that long though, its pretty bad.

My opinion, stop it, clone it with linux like dd_rescue, make a full clone if possible, then take that clone and use something like R-studio or Getdataback and scan it, save what you can that the CHKDSK didnt fuck up.

 

[BadOmen]

I would force explorer to show all files, and not hide anything, then check what the files are.

Did that, but it showed nothing. No .chk files, no found.xxx folder.

So I booted on my Samsung HDD and ran Recuva on the crashed Seagate disk. It managed to recover almost everything, apparently (funny, Recuva marked them as non-deleted). Files just lost all their folder structure, but I guess it's all there (plus, my backups weren't that bad either). In terms of data, I think I'm ok. MASSIVE PHEW!

Due to that and other reasons, I'm reformatting the SSD, now with no SATA port swap after. I've already installed Windows and updated all drivers, will connect the storage drives tonight, format the partition that crashed and see what happens. Elixer, I'll also make sure crystaldiskinfo sniff around this drive.

I'm curious if the Barracuda 3TB could have bad blocks (although it's too new for that, come on, Seagate!). Guess a regular format will tell me that, right?

 

[Elixer]

Did that, but it showed nothing. No .chk files, no found.xxx folder.

So I booted on my Samsung HDD and ran Recuva on the crashed Seagate disk. It managed to recover almost everything, apparently (funny, Recuva marked them as non-deleted). Files just lost all their folder structure, but I guess it's all there (plus, my backups weren't that bad either). In terms of data, I think I'm ok. MASSIVE PHEW!

Due to that and other reasons, I'm reformatting the SSD, now with no SATA port swap after. I've already installed Windows and updated all drivers, will connect the storage drives tonight, format the partition that crashed and see what happens. Elixer, I'll also make sure crystaldiskinfo sniff around this drive.

I'm curious if the Barracuda 3TB could have bad blocks (although it's too new for that, come on, Seagate!). Guess a regular format will tell me that, right?

No, once you format the HD, then it clears the bad blocks (remaps them) so you aren't going to know.
This is why you look at it with CrystalDiskInfo first, and see what it shows.

Congrats on the files save though, wish everyone was that lucky.

 

[Cerb]

(edit: CHKDSK running on ANY drive is bad news, if there is important data on a drive, NEVER run chkdsk if it wants to, copy your data out using 3rd party utility, etc before you run it.)

TBF, I run w/o a UPS, and CHKDSK sometimes has do a full pass after a bad storm.

Normally, it takes 1-1.5hrs for my 1TB. I would have stopped it if it weren't making lots of progress within 2 hours on a 3TB.

Glad the OP got some of his stuff.

Now, backup, backup, backup!

 

[BadOmen]

I would have stopped it if it weren't making lots of progress within 2 hours on a 3TB.

I could see it wasn't stuck on some percentage count, in fact it was trying to recover files, one after another, but had no space to do so and thus taking a lot of time per file recover attempted.

But I don't get what triggered the whole chkdsk thing. Files were properly recovered in the end, nothing wrong with them, and I checked Crystaldiskinfo and it says the drive is all good, all SMART variables ok.

To be on the safe side, I'm not plugging stuff to Marvell-controlled ports again. So long for SATA 3.
I thank you all for the help!

 

[murphyc]

Guess a regular format will tell me that, right?

A regular format zero's the drive. Upon writing anything, the drive firmware immediately confirms the write, and if it fails it'll try again. If the write failure is persistent, the sector is removed from use by the firmware. In any case with a modern drive, a regular NTFS format should not record any bad blocks. The drive firmware deals with it.

Before firmware dealt with this, file systems had to, hence the antiquated feature in NTFS for dealing with this. On Linux, ext also has such a feature.

You're better off doing an ATA Secure Erase on both HDD and SSD because it's faster than zeroing, and safe for SSD whereas writing zeros causes unecessary SSD wear. Then do fast formatting (which at least for installing Windows 7 happens to be the default, I'm not sure if it is for additional drives.)

 

[Piotrsama]

(Sorry if mentioned, didn't read the whole thread)

Your 3TB HDD has a single partition with all the space allocated?
Could it be that you recently passed the 2.2TB used capacity? (there's been some problems reported about that).

 

[Dessert Tears]

It fixed index problems, and now it's going file by file with the messages:
- copying orphaned file ____.__ from directory xxxx to xxxx (or something like that)
- insufficient disk space to recover lost data

I had a similar incident nearly 10 years ago involving Windows 2000, the 127GB limit, and a drive overlay (never use one). When I shut down Windows, chkdsk started running with similar messages for each file. I turned the computer off after a few minutes, after it had overwritten the first 100MB of the drive. The later partitions were fine after I used a pay tool to fix the MBR. I wrote off the first partition: low value files and pretty high fragmentation < $80 for GetDataBack. I'm probably going to try again with Recuva before I shred and get rid of the drive.

 

[BadOmen]

That's unbelievable. So I formatted only the partition that was affected (this Seagate 3TB has two partitions), and after that, started copying files from the backup to it.

After restarting the computer, the whole hard drive is gone. Disk Management sees it as 3TB "unnalocated" and asks if I want MBR or GPT.

Now I need to find out:
- How to recover data in such situation? Not fundamental, but I'd like to salvage a few things from the other partition.
- Is the Mushkin Chronos SSD having compatibility issues with the Barracuda 3TB?
- Is the disk dying? Although it's quite new and CrystalDiskInfo didn't raise any flags while it was working.

WEIRD UPDATE: Windows installed in the SSD sees the 3TB as unallocated AND only 746.52 now (via Disk Management). Windows at my old Samsung HDD sees it as 3TB.

 

[murphyc]

Windows installed in the SSD sees the 3TB as unallocated AND only 746.52 now (via Disk Management). Windows at my old Samsung HDD sees it as 3TB.

It's either a hardware problem, or there's corruption, or both. Those things explain intermittent problems that seem random. The GUI will often simplify different partition states such that they are graphically described the same way. So you don't really know what the state is. This is the same obscurity issue on OS X and Windows, so really the only thing to do is go to the command line to get the truth.

For GPT, my preference is GPT fdisk, also known as gdisk. There is a Windows binary. I suggest using that, but you'll need to check the documentation for the Windows specific syntax as I don't know how disks are specified on Windows vs partitions. On OS X and Linux it would be something like:

sudo gdisk -l /dev/disk0

 

[BadOmen]

For GPT, my preference is GPT fdisk, also known as gdisk.

It's definitely worth a try, thanks. By using Minitool Partition Wizard, I managed to save the partition I had just formatted. The other one, the one that has never had anything to do with the story, is still unallocated.

The big mystery for me is: why does only the SSD keep messing with the Barracuda 3TB? When I boot from my old HD instead, everything works fine with it.

 

[murphyc]

The big mystery for me is: why does only the SSD keep messing with the Barracuda 3TB? When I boot from my old HD instead, everything works fine with it.

It's unlikely it's the SSD doing anything. It's more likely the software or settings on the SSD that differs from the old hard drive. I'm a little hard pressed to imagine how one drive could cause corruption on another, other than a signaling problem. I'd check all the usual suspects and make sure the firmware is current for the computer, the HBA, and all the drives.

 

[BadOmen]

It's more likely the software or settings on the SSD that differs from the old hard drive.

You were right. I think I've found the culprit.
Turns out newer Intel RST drivers are not known for working well with P55 boards. And older drivers don't support 2.2TB+ HDDs. So, when I bought this 3TB Seagate, I installed DiskWizard to unlock its full capacity.

Once I reinstalled Windows on the new SSD, I installed the latest Asus-approved version of the Intel drivers, which is old Matrix Storage with no 2.2TB+ support. AND FORGOT DiskWizard. Nice.....

May my sad (and dumb) experience be registered here as a word of caution for everybody else playing with large HDDs in older boards.

And my big thank you to everybody who posted in this thread.

 

[Dessert Tears]

May my sad (and dumb) experience be registered here as a word of caution for everybody else playing with large HDDs in older boards.

I read a bit about DiscWizard and "Beyond 2TB". Google results mention older versions of DiscWizard and Dynamic Drive Overlay, and it seems like the latest one also includes DDO. Based on my experience with a Western Digital DDO and chkdsk eating my data, I recommend that you find another solution, such as a controller card or using only the first 2.2TB.

Edit: I re-read your posts and realized that you were running fine long enough to forget that you had installed DiscWizard. It was only after you installed the SSD that you had problems. Mine came out of the blue, as I recall.