Reverting to physical disk after importing with qm import-from

[fmaione]

Hello, I just imported a physical disk to a VM using qm set 998 -scsi0 local-storage-zfs:0,import-from=/dev/sdd

There is a way to do the opposite, since the physical disk is the same? As the name says, VM disk is stored on ZFS.

Many thanks

 

[bbgeek17]

You can try to pass-through a disk to VM, then use "disk>move" in VM hardware option. I am not sure if it will work.
You can always use "qemu-img" directly to move image around, take a look at "man qemu-img" or search for "qemu-img examples".

If the disk is stored in ZFS in raw format, you may even be able to use simple "dd".

Good luck

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[fmaione]

@bbgeek17 you were absolutely right. Although virtual disk had a bunch of KB more than physical one, I just reverted using dd. It was a ZFS pool stored on the two-partitions layout (sdd1 and sdd9) and despite dd final error for going short writing on physical device, zpool scrub went absolutely well

The full story:

1. I imported a physical disc to a VM using qm set 998 -scsi0 local-storage-zfs:0,import-from=/dev/sdd
2. I used it
3. I reverted to physical disk doing:
   • attached the physical disk via pass-through to the vm using qm set VMID -scsiX /dev/disk/by-id/ata-SanDisk_SDSSDA240G_17427xxxxxx,vendor=SanDisk,product=SDSSDA240G,serial=17427xxxxxx,cache=writeback,iothread=1,backup=0,replicate=0,rerror=report,werror=report
   • executing dd if=/dev/VIRTUALDISK of=/dev/PHYSICALDISK bs=10M status=progress

So if you get an error on the final stage of write it's because virtual disk has too much sectors.

Use at your risk, make sure you have backups.