ZFS Pool problem

[Antonius]

Hello,

I need some help, I got a pool on TrueNas Core.

After using it i wanted to update to TrueNas Scale.
I am running TrueNAS on a esxi node.

I always make snapshots for upgrades etc, but!
This time during the upgrade the disk where TrueNAS OS / Disks where on was full.
Meaning that ESXI.

At this moment TrueNas lost sight of my pool.
When trying to import it i get Cannot import 'POOL': I/O error


When i try to import it with -f i get WARNING: can`t open objset 270, error 5

The object sets are 270,262,387,278,270,278,262,387,270,387,278,262 in that order.

I really need to get access to the data on this pool so i hope one of you guys can help me.
Even if i need to do data recovery and recreate the pool, as at least i can get a backup from the data.

Pool is in stripe mode.

 

[jgreco]

Welcome to the forums.

Sorry to hear you're having trouble. Please take a few moments to review the Forum Rules, conveniently linked at the top of every page in red, and pay particular attention to the section on how to formulate a useful problem report, especially including a detailed description of your physical virtualization hardware, along with a description of your virtual hardware, what your PCIe passthru configuration is, and what sort of controller you've used for passthru.

You've basically given no one anything to work with, so the responses will tend to be random guesses rather than anything useful.

 

[HoneyBadger]

This time during the upgrade the disk where TrueNAS OS / Disks where on was full.
Meaning that ESXI.

I have some concern here that you assigned virtual disks to TrueNAS, and the underlying VMFS datastore became full at the time you attempted to upgrade.

Further guessing is that the virtual disks assigned to TrueNAS were thin-provisioned, such as that they could potentially have outgrown the physical disks backing them. This would have caused your TrueNAS VM to pause/stun itself when the backing disk reported the ENOSPC/"unable to grow" error.

Is this an accurate assumption?

 

[Antonius]

The disks where thick provisioned, but yes i did get a pause/stun because they where thick provisioned as they will claime al l the space.
I am using

VMWare ESXI
2 disk`s are direct passthru
2 disk`s are added as VM storage

 

[jgreco]

Further guessing is that the virtual disks assigned to TrueNAS were thin-provisioned, such as that they could potentially have outgrown the physical disks backing them.

This is a questionable conclusion; a thick provisioned virtual disk can also easily outgrow the physical disks when snapshotted, etc., though it is obviously easier for it to happen on thin-provisioned especially with a CoW filesystem.

This is just one of the reasons that real physical disks need to be passed through to TrueNAS via PCIe passthru.