[SOLVED] (6.8.0) ISO for Win10 VM on Unassigned Device died seemingly out of nowhere

[TurkeyPerson]

Hey all,

 

Basically, I restarted my server yesterday and when I tried to boot my windows 10 VM, I got the oxc00000f error from windows. Attempts to repair indicate it can't find a windows installation. Looking at the drive it's running on which is an unassigned device (not passed through), an NVME drive, I saw the following errors in the log (repeating) every time I load the VM:

 

Tower kernel: print_req_error: critical medium error, dev nvme0n1, sector 30345600

Tower kernel: Buffer I/O error on dev nvme0n1p1, logical block 3792944, async page read

 

I did some research, and that lead me to believe maybe the SSD was dying - seemed unusual, as it's not that old - but possible (it's a Force MP500). I did some testing, copied a file onto it and off of it - everything seemed fine.

When I tried to copy the ISO using dolphin it fails because it "cannot read" it. I assume this is some kind of file corruption? Any tips/fixes? Even how to avoid this in the future. I know I should have taken back ups - so I'll be doing that going forward, but I'm really hoping not to lose everything.

Note: I'm currently on 6.8 because I downgraded to test whether it could help - it didn't. I was on 6.8.2 to begin with.
 

Thanks for reading!

tower-diagnostics-20200507-0955.zip

Edited July 28, 2020 by TurkeyPerson
Attaching diagnostic properly

[trurl]

Attach your Diagnostics instead of linking to external sites. Thanks 

[TurkeyPerson]

Hey! Thanks for the reply! Sorry about that, I didn't see the option in the original submission. Here you go.

tower-diagnostics-20200507-0955.zip

[JorgeB]

NTFS filesystem corruption, run chkdsk.

[JorgeB]

Also, would recommended using a Linux native fs for the unassigned device that contains the vdisk, any special reason you're using NTFS?

[TurkeyPerson]

27 minutes ago, johnnie.black said:

NTFS filesystem corruption, run chkdsk.

When I run "wmic logicaldisk get caption" through windows install disk troubleshooter via CLI, I get three drives D, E, and X.
It seems like D is the install disk (UDF?), and E is "RAW" and I can't run chkdsk on it, and finally x which I assume is just a live disk of some kind is write protected. So, E is the drive and the partition is gone? 

If that's the case, I assume I need to use some kind of recovery software? Does that mean building a new VM with something like EaseUS partition recovery and adding the drive to that VM as a secondary?

EDIT: diskpart > list disk tells me that there are no fixed disks to show.

EDIT2: Alright so list volume shows me VirtIO and what I assume is the windows install disk. So I guess it's not even seeing the drive.

17 minutes ago, johnnie.black said:

Also, would recommended using a Linux native fs for the unassigned device that contains the vdisk, any special reason you're using NTFS?

I just didn't know any better. Thank you i'll look into that.

Thank you for both replies!

Edited May 7, 2020 by TurkeyPerson

[JorgeB]

You'd need to run chkdsk on the unassigned device itself, not using the VM, connect the device on a Windows desktop, or boot your server with a Win10 install flashdrive and use the command line.

[TurkeyPerson]

Oh! I see. Okay I'll try to boot the server with a windows 10 drive, thank you.