[OpenIndiana-discuss] ZFS Error

[Joshua M. Clulow]

Hi Gaetano,

I'm sorry to hear about your drive troubles.  Data loss is always very
unfortunate.  I've put some questions that you could answer for us
towards the end of the e-mail, which might help us investigate what
happened to your drive.

On Fri, 22 Jan 2021 at 21:36, Hung Nguyen Gia via openindiana-discuss
<openindiana-discuss at openindiana.org> wrote:
> It's from experience. If a pool faulty on FreeBSD, I will try to import it on Linux or OI.
> Of course, I will try to import it from FreeBSD first.
> But most of the time it doesn't work.
> The key is to import on other OSes that support ZFS.
> The chance that it will be able to import the pool is higher.

This is not good advice.  It sounds like there is corruption in the
pool.  Making random changes to that software you're using (e.g.,
different versions of the ZFS, or even different operating systems)
doesn't increase the chance that you'll be able to import the pool; it
merely increases the risk that you will make things worse.

With corrupt file systems, you're generally in what is often described
as an "incident pit":

    https://en.wikipedia.org/wiki/Incident_pit

Each time you make a move, you can either climb out (by fixing
something) or fall further in (by making a bad choice).  Doing random
things to the disk is very likely to drive you further into the pit;
e.g., even importing a pool is a process that tries to write to the
(possibly fragile or corrupt) media, and may make things worse.

If you genuinely care about the data on the device, I would first make
an image of the whole device and store it as a backup, using a
read-only tool like "dd".  Then you can continue to operate on the
errant volume with much lower risk.

If you don't care about the data you've stored on the pool, and are
asking why it has become corrupt after so little activity, we'll
likely need some more information.  Some USB storage devices are
notoriously unreliable, especially with respect to I/O operations that
are critical to zpool integrity like cache flushes.

~ ~ ~ QUESTIONS: ~ ~ ~

Some questions that might provide useful background information, either way:

  - Is it a sealed all-in-one USB SSD unit, or an SSD device sitting
in a USB enclosure?

  - What model enclosure and/or SSD device?

  - How old is the device, and have you used it in other systems before?

  - Before the data loss event, did the system panic, or was the power
to the system interrupted, or did the USB cable become disconnected at
any point?


Cheers.

-- 
Joshua M. Clulow
http://blog.sysmgr.org