Hi Unraid team,

I’d like to suggest a feature for future ZFS support in Unraid. Currently, ZFS pools are typically created using Linux device names like sda, sdb, etc. While this works, it can cause issues if disks are moved between systems, controllers, or after hardware changes, since device names are not guaranteed to stay the same.

I guess that It would be helpful to have an option to create ZFS pools using persistent device IDs (e.g., /dev/disk/by-id/...). This would:

• Make pools more portable and robust across hardware changes

• Reduce the risk of import or resilvering issues caused by device renaming

• Improve reliability for NAS setups with multiple disks

This could be offered as an optional toggle during pool creation, so existing workflows are not affected.

However, it’s worth considering potential downsides, that actually I don't know. Additionally, there could be compatibility concerns with certain hardware or Unraid’s current implementation.

What do you think about it?

11 minutes ago, skler said:

it can cause issues if disks are moved between systems, controllers, or after hardware changes, since device names are not guaranteed to stay the same.

Unraid always uses the current names, since the devices have to be assigned.

Maybe I’m wrong, but to expand a pool I’ve used the device ID instead of the Linux device name, and it seems to work:

 zpool status storage
  pool: storage
 state: ONLINE
  scan: scrub repaired 1000K in 1 days 00:03:24 with 0 errors on Mon Oct 13 03:30:25 2025
expand: expansion of raidz2-0 in progress since Mon Oct 13 10:49:25 2025
	65.2G / 87.8T copied at 642M/s, 0.07% done, 1 days 15:48:19 to go
config:
	NAME                                  STATE     READ WRITE CKSUM
	storage                               ONLINE       0     0     0
	  raidz2-0                            ONLINE       0     0     0
	    sdk1                              ONLINE       0     0     0
	    sdn1                              ONLINE       0     0     0
	    sdr1                              ONLINE       0     0     0
	    sds1                              ONLINE       0     0     0
	    sdt1                              ONLINE       0     0     0
	    sdh1                              ONLINE       0     0     0
	    ata-ST20000NE000-XXXXXX_XXXXXXXX  ONLINE       0     0     0
errors: No known data errors

My question is: if I move, for example, the disk currently in sdk to another slot, or to another controller (e.g., in a JBOD), will the pool still recognize the same disk based on its device name, or would I need to reference its ID again?

Edited October 13, 2025Oct 13 by skler

I would recommend using the GUI to expand the pool, but as long as the device is assigned to the pool, Unraid will use it. It doesn't matter what its current ID is; whatever it is, it will be used during pool import.

Ok, got it. I’m still on 7.1, once 7.2 is released as stable, I’ll upgrade. In the meantime, I’ll use the same approach as the GUI from the CLI (create a new partition and add it by the Linux device name).

Thanks a lot!