0

I'm currently very desperate: I have a 3TB WD Red disk that I need to extract data from.

This was part of a softraid on an OpenBSD 5.6 machine. I was in the process of migrating to a zfs on linux machine, and all hell has broken loose:

  • The original machine now fails to POST; no chance of going back into the original install.
  • I had booted into an OpenBSD shell, with both disks attached (it could see everything fine, no probs) when one of the disks literally caught on fire - I doubt it's going to be usable again, the PCB has melted.
  • My external offsite backup got corrupt too - all data lost on it.

So I'm now sitting with my one copy of data on a machine that cannot have OpenBSD installed on it (newer ASRock motherboards seem to have a UEFI fit with an OpenBSD formatted disk), leaving me needing to boot from CD and drop to a shell, and attempting to backup to a usb disk from there.

I've tried mounting the disk with the data on but it fails with an I/O error. A disklabel sd0 sees the disk fine, but with a type of 'RAID' - when I had previously booted (before the fire) OpenBSD had seen it as a good softraid setup and mounted it successfully.

Any ideas what I can do from here? I have many years of data on the disk (including current Open University details and all my old software projects) - the OpenBSD man pages have brief detail and I don't want to risk running a command that might destroy/rebuild the array.

ZXcvbnM
  • 111
  • 4

1 Answers1

0

Luckily, before I attempted this migration I had sync'd the softraid with another individual disk. While this wasn't able to be mounted within a boot shell, another OpenBSD install I had was able to read the disk without an issue.

I'm currently copying the 2TB of data over the network into a single-disk zfs which I shall upgrade to a mirror (using the raided disks) once I can confirm it's all there.

I suspect I could use bioctl -R to rebuild the array from this other box, but I couldn't initialize another disk within the CD boot shell, and didn't want to alter the other box as best as possible. All the power on/offs I've been doing also increases the risk of failure, and with the luck I've had of late I didn't want to push it.

ZXcvbnM
  • 111
  • 4