Moving from ESXI to PVE with an existing ZFS pool

[attley]

Hello,

I installed PVE and hoping to get my old ZFS pool (used with FreeBSD ESXI VM) imported to PVE to connect as a storage to a container which then would share the ZFS via an NFS/SMB service to my home network. From the posts I have read online, I assume this is possible. The problem is that when the SATA drives are connected to the PVE system, the system cant seem to find any existing zfs pools. Is there some special configuration that is needed here?

root@pve:~# zpool import
no pools available to import

root@pve:~# zpool list
no pools available

The discs show up in the following way:

root@pve:~# lsblk
NAME                    MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
...
sdb                       8:16   0   1.8T  0 disk
└─sdb1                    8:17   0   1.8T  0 part
sdc                       8:32   0   1.8T  0 disk
└─sdc1                    8:33   0   1.8T  0 part

One thing I noticed, that the discs are showing as "VMFS_volume_member". If this is related to taking the discs out of ESXI system, do I need to perform some partition "cleaning" or wiping some ESXI related info from the discs in order to make them recognized and how would one achieve that?




Cheers,
Olli

 

[apoc]

One thing I noticed, that the discs are showing as "VMFS_volume_member". If this is related to taking the discs out of ESXI system, do I need to perform some partition "cleaning" or wiping some ESXI related info from the discs in order to make them recognized and how would one achieve that?

This indicates that you have used VMFS on your esxi which I was about to ask ( and which is a common setup).

What you try to do in the way you do is not possible, as VMFS is a proprietary file system. PVE just can't read it. And AFAIK there is no 3rd party driver for Linux...

What you would need to do is get them back in esxi and pull the vmdk files including their structure (flat and descriptor) to your PVE, then convert these disks to a PVE readable format.

Do you have spare capacity to achieve that?

 

[apoc]

Important: do not modify the disks! Dont delete partitiond etc. You will screw (or might) your data !!!

 

[attley]

This indicates that you have used VMFS on your esxi which I was about to ask ( and which is a common setup).

What you try to do in the way you do is not possible, as VMFS is a proprietary file system. PVE just can't read it. And AFAIK there is no 3rd party driver for Linux...

What you would need to do is get them back in esxi and pull the vmdk files including their structure (flat and descriptor) to your PVE, then convert these disks to a PVE readable format.

Do you have spare capacity to achieve that?

Ok. Unfortunately I dont have the spare capacity. But as a B plan I do have backups for the important data of the mirror, so I if there is no other way, I might just wipe the 2x2TB discs and create a whole new ZFS pool and the rsync my data to the new pool? Not the nicest way but this could be done.

 

[UdoB]

And AFAIK there is no 3rd party driver for Linux...

There is this:

vmfs-tools/stable 0.2.5-1+b2 amd64
  Tools to access VMFS filesystems

vmfs6-tools/stable 0.1.0-3 amd64
  Tools to access VMFS6 filesystems

So file-recovery should be possible without re-instantianting a complete ESXi. Probably those tools are NOT sufficient to run vmfs in a "productive" setting with good performance.

Best regards

 

[apoc]

There is this:

Cool. Thanks.

But still you need to get the data off the system...