HDD Wipe Crashes FreeNAS, Now ZFS Volume Won't Import

[Gravypicante]

Hey everyone, I'm having trouble re-importing a ZFS volume that appears to be intact. Any help or insight would be greatly appreciated.

Build: FreeNAS-9.2.1.8-RELEASE-x86
Platform: Intel(R) Pentium(R) 4 CPU 3.20GHz
Memory: 2017MB​

The volume started acting weird after FreeNAS unexpectedly crashed (rebooted). I had used the GUI to wipe a disk so that I could add it to a different ZFS mirror. That disk has since been added to the unrelated mirror without problems, but clicking the "Wipe" button in the GUI is what caused the first crash.

The ZFS mirror I am having trouble with consists of two disks, connected with USB. Attempting to restart FreeNAS after the first crash always ended with a "Fatal trap 12: page fault while in kernel mode" error. So I detached the volume after successfully restarting FreeNAS with both disks disconnected.

Now FreeNAS itself is working fine, and so is my other ZFS mirror (which consists of 3 SATA disks). But any attempt to re-import my first ZFS mirror instantly crashes FreeNAS and the computer reboots. Using "zpool import" shows that the detached volume appears to be fine:

# zpool import
pool: SYF-Files
id: 12499899812707172444
state: ONLINE
action: The pool can be imported using its name or numeric identifier.
config:

SYF-Files ONLINE
mirror-0 ONLINE
gptid/a20fbc90-63cb-11e4-9a62-00123f3a71a4 ONLINE
gptid/b001b80c-63cb-11e4-9a62-00123f3a71a4 ONLINE​

I suppose there is a log somewhere that would show what exactly is happening, but I'm not sure which log would be most helpful (or where I would find it). Suggestions?

I've tried the following methods of re-importing the volume:

• GUI feature "Auto Import Volume"
• Regular zpool import command
• Zpool import command with force option
• Zpool import command with discard unfinished changes option

I would simply remake the mirror if I had a recent backup of the data. It's probably not surprising that I wasn't smart enough to do that and my only backup is several months old.

So yes, this is mostly my fault, but I'd still be infinitely grateful if someone could help me get things up and running again!

 

[Bidule0hm]

You only have 2 GB of memory, the minimum requirement is 8 GB.

 

[Gravypicante]

You only have 2 GB of memory, the minimum requirement is 8 GB.

It actually had 1GB for some time. I know that 2GB is below the recommended amount. However, the graphs under "Reporting" show that FreeNAS seems to be content using less than half of the available memory. This is just a home server by the way; no more than 4 or 5 computers are using the server at any given time.

Do you really think that FreeNAS is crashing because it's running out of memory? Is there an easy way to confirm this?

Unfortunately, the old piece-of-junk computer that FreeNAS is running on can't take more than 2GB of memory. I'll do whatever is necessary to get my data back, but I'd like to make sure that building a new server with 8GB of memory would actually solve the problem.

 

[Bidule0hm]

This is because you have so not enough RAM that it is misused.

It can be a home server for only one user the requirement is still 8 GB minimum. How to confirm? just read the dozens of threads with crashing servers that doesn't meet the requirements...

"I'd like to make sure that building a new server with 8GB of memory would actually solve the problem." It will. If you follow the other requirements too of course (8 GB of RAM, no hardware RAID card, ECC RAM, server grade MB, ...), just read the hardware recommendations thread (link is in my signature) ;)

 

[Gravypicante]

This is because you have so not enough RAM that it is misused.

It can be a home server for only one user the requirement is still 8 GB minimum. How to confirm? just read the dozens of threads with crashing servers that doesn't meet the requirements...

"I'd like to make sure that building a new server with 8GB of memory would actually solve the problem." It will. If you follow the other requirements too of course (8 GB of RAM, no hardware RAID card, ECC RAM, server grade MB, ...), just read the hardware recommendations thread (link is in my signature) ;)

I suppose that would explain why two seemingly unrelated actions both crash the server (wiping a disk, and trying to import a volume).

Thanks for the response. I'll trust your advice and see what hardware I have access to for building a new server.

... and I did spot the
Hardware recommendations (read this first)
thread, so I'll be sure to check that out.

 

[krikboh]

The pool is most likely corrupted and unrecoverable from running with such low memory. New hardware will not recover the data.


Sent from my iPhone using Tapatalk

 

[SweetAndLow]

The pool is most likely corrupted and unrecoverable from running with such low memory. New hardware will not recover the data.


Sent from my iPhone using Tapatalk

I agree with this almost never do people recover a pool that can't import because of low memory. I've seen 1 successful recovery out of 50.

 

[Gravypicante]

The pool is most likely corrupted and unrecoverable from running with such low memory. New hardware will not recover the data.


Sent from my iPhone using Tapatalk

I assumed that the pool might be corrupted (the configuration/metadata whatever). But I was hoping that the actual data was fine and could be accessed somehow. Is that a fantasy? Is the data itself not independent from the pool setup?

Are there any utilities for repairing broken pools that I missed? I already tried zpool import with the discard unfinished changes option.

 

[SweetAndLow]

The os configuration is separate from your pool. But your pool is corrupt so not much you can do. There are no recovery tools for zfs unless you want to pay 5 figures.