Can I do this? (Clone a failed disk in a stripe pool)

TrueNAS General
1

Hello,
I have a stripe of 5 HDDs, and 1 disk failed. I couldn’t get the pool back without the disk. I took the disk to a recovery company, and they said it has an ABR issue. They asked if we can clone the HDD on another working 10 TB HDD. If I get it cloned and put the new HDD back on the stripe, will I be able to get the pool back?

2

Hi and sorry to hear this.

I honestly don’t know is the answer, I think it would depend on many things.

I mean this in the nicest possible way but why have a stripe of 5 disks if you care about the data you have on the pool? Perhaps this was created in error?

3

Yeah, it was created by a person who had no knowledge of the RAIDs.
Am trying to find a solution for this.

4

Fair enough. Like I say I honestly don’t know but if the data is valuable then it’s perhaps worth a shot.

5

The clone might work, but you’re in “hail Mary” territory.

Out of curiosity, what is “ABR” in this context?
I assume it is none of American Board of Radiologists, Applied Biological Research, American Business Review, Auditory Brainstem Response, or Antibiotic Resistance.

1 Like
6

If the drive is not working before cloning, then why do they / you think you can read enough of it to clone it properly?
And good enough for ZFS to allow data recovery?

Some recovery services can recover data off hard drive platters such that they can clone the data to a new hard disk drive. Then, they return the new hard disk drive, (and possibly the broken one), to you. It sounds like that recovery service you contacted is not one that works at the platter level.

That said, I don’t know what “ABR” means in this context either…

For those that don’t know:

  • ZFS redundancy is at the vDev level, not the pool level
  • Single disk vDevs have no redundancy
  • ZFS will normally stripe data across all vDevs

The results of this is likely corruption of all files large enough to go across all vDevs. Meaning good data still exist on the good disks, (and good blocks on the bad disk), but larger files will could have holes in them from a bad disk. Which in some file types makes them useless.


Now that I have said all that, their was a ZFS project years ago to allow importing a pool with missing vDevs. Then allow extraction of any good data. Don’t know status of that project, nor if it was ever finished. And definitely not how to activate that mode, (if it exists).

Further, that is EXTREME recovery and beyond my skill set. So don’t ask me about it. If you want / need to pursue it, someone else will have to help. Or check the OpenZFS documentation.