Disk Names in zpool status (lack of /etc/zfs/vdev_id.conf support)?

TrueNAS General
, ,
1

My disk names look like this:

  pool: fast
 state: ONLINE
  scan: resilvered 1.79M in 00:00:00 with 0 errors on Sun May 11 22:41:04 2025
config:

        NAME                                      STATE     READ WRITE CKSUM
        fast                                      ONLINE       0     0     0
          mirror-0                                ONLINE       0     0     0
            1edb3237-5414-4b26-9332-5a74b9d2d13c  ONLINE       0     0     0
            8e8826fa-44a5-4814-956a-ee01401226f6  ONLINE       0     0     0
          mirror-1                                ONLINE       0     0     0
            a48011cc-aca2-401f-8294-37bc8e9e9d6f  ONLINE       0     0     0
            2678bfb0-e91b-4510-9bc9-a703fedd48c6  ONLINE       0     0     0

errors: No known data errors

  pool: rust
 state: ONLINE
config:

        NAME                                      STATE     READ WRITE CKSUM
        rust                                      ONLINE       0     0     0
          raidz2-0                                ONLINE       0     0     0
            1b3cd88f-4a12-4ae6-8b04-05318c8519cf  ONLINE       0     0     0
            8f4ce5b1-4ca5-4eab-b3b5-9b800cf782ab  ONLINE       0     0     0
            eba7bd0f-4b75-4d42-84fb-eb75eee16a88  ONLINE       0     0     0
            fc870b89-2592-4066-96c6-5e4897f1a168  ONLINE       0     0     0
            12404dd4-4d87-4038-838c-c60796116136  ONLINE       0     0     0
            931437b1-01d1-4ea6-aea0-9993b465fa36  ONLINE       0     0     0
        cache
          d4cb4922-db0e-4298-8f87-5a5e0eef9d45    ONLINE       0     0     0

i would prefer them to refelct the drives (to make it easier when i have an issue and will be running around with my hair on fire).

yes i know can export the pool and import by-id the issue with that is there are multiple symlinks per drive in /dev/disk/by-id and it sometimes picks silly non-useful ones

i went down the path of /dev/disk/by-vdev with a /etc/zfs/vdev_id.conf route

but it seems that truenas doesn’t ship with the required tools or create the require symlinks

is there a good reason why truenas doesn’t support this?
is there a good alternative workaround?

2

Try sudo zpool status -LP.

And yes, there is a (good) reason why TrueNAS uses partuuids when defining vDev members.

3

cool, i would love to learn what those are

4

well i did that before i posted, its no better, as these names are dynamic and can change each boot on linux this is why a hard mapping (using the file) between uuid/guid and a friendly name is prefered and recommened by upstream zfs folks.

i will see if there is or add a feature request

 state: ONLINE
  scan: resilvered 1.79M in 00:00:00 with 0 errors on Sun May 11 22:41:04 2025
config:

        NAME                 STATE     READ WRITE CKSUM
        fast                 ONLINE       0     0     0
          mirror-0           ONLINE       0     0     0
            /dev/sdh1        ONLINE       0     0     0
            /dev/sdi1        ONLINE       0     0     0
          mirror-1           ONLINE       0     0     0
            /dev/nvme10n1p1  ONLINE       0     0     0
            /dev/nvme8n1p1   ONLINE       0     0     0

errors: No known data errors

  pool: rust
 state: ONLINE
config:

        NAME              STATE     READ WRITE CKSUM
        rust              ONLINE       0     0     0
          raidz2-0        ONLINE       0     0     0
            /dev/sdb1     ONLINE       0     0     0
            /dev/sdc1     ONLINE       0     0     0
            /dev/sdd1     ONLINE       0     0     0
            /dev/sde1     ONLINE       0     0     0
            /dev/sdf1     ONLINE       0     0     0
            /dev/sdj1     ONLINE       0     0     0
        cache
          /dev/nvme2n1p1  ONLINE       0     0     0
6

Looks like you answered your own question… :tada:

7

not really, as neither of these have been answered

is there a good reason why truenas doesn’t support this?
is there a good alternative workaround

(except maybe the last one is ‘no’ ?)

8

Instead of trying to change the default behaviour of the system why not use something like this to make better sense of the system when its disk identification time.

lsblk -o NAME,model,tran,size,serial,label,partuuid,path | sed -e "`ls -1cd /sys/class/enclosure/*/*/device/block/*|sed "s+.*enclosure/\(.*\)/device/block/\(.*\)+s-\2\\$-\2 \1-+"`" | sort -t ':' -k 3
9

thanks for the useful suggestion, that gives me some paths to explore!

10

because i would have no idea how to write an awk statement, or do any coding… and no idea the enclosure path existed

but you made a nice suggestions and with some chatgpt assisted coding later and i now have a script that outputs this (blank zfs details means disk not yet used in ZFS) - so many thanks!

NAME       MODEL                        TRAN   SIZE     SERIAL               PATH               POOL       TYPE       GUID                                
sdb        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sdb           rust       'disk'     15092755422561716895                
sdc        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sdc           rust       'disk'     4022156489591843608                 
sdd        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sdd           rust       'disk'     16272000233204181760                
sde        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sde           rust       'disk'     2296675024925190545                 
sdf        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sdf           rust       'disk'     1043860037485403250                 
sdg        ST4000VN000-1H4168           sata   3.6T     xxxxxxxx             /dev/sdg                                                                     
sdh        ST4000VN000-1H4168           sata   3.6T     xxxxxxxx             /dev/sdh                          
sdi        ST4000VN000-1H4168           sata   3.6T     xxxxxxxx             /dev/sdi                          
sdj        ST24000NT002-3N1101          sata   21.8T    xxxxxxxx             /dev/sdj           rust       'disk'     10764108644732859737                
nvme4n1    INTEL SSDPE21D015TA          nvme   1.4T     xxxxxxxx             /dev/nvme4n1                                                                 
nvme2n1    ADATA SX8200PNP              nvme   1.9T     xxxxxxxx             /dev/nvme2n1                                                                 
nvme6n1    INTEL SSDPE21D015TA          nvme   1.4T     xxxxxxxx             /dev/nvme6n1                                                                 
nvme0n1    KINGSTON SEDC2000BM8960G     nvme   894.3G   xxxxxxxx             /dev/nvme0n1                                                                 
nvme5n1    INTEL SSDPE21D960GA          nvme   894.3G   xxxxxxxx             /dev/nvme5n1                                                                 
nvme7n1    Seagate ZP4000GM30063        nvme   3.6T     xxxxxxxx             /dev/nvme7n1                                                                 
nvme3n1    INTEL SSDPE21D960GA          nvme   894.3G   xxxxxxxx             /dev/nvme3n1                                                                 
nvme10n1   Seagate ZP4000GM30063        nvme   3.6T     xxxxxxxx             /dev/nvme10n1      fast       'disk'     11559379796082310176              
nvme9n1    Seagate ZP4000GM30063        nvme   3.6T     xxxxxxxx             /dev/nvme9n1                                                                 
nvme1n1    KINGSTON SEDC2000BM8960G     nvme   894.3G   xxxxxxxx             /dev/nvme1n1                                                                 
nvme8n1    Seagate ZP4000GM30063        nvme   3.6T     xxxxxxxx             /dev/nvme8n1       fast       'disk'    10845657097477000304

as to why i would ask for a standard ZFS feature to be on a system using standard ZFS - that would seem an obvious, simple, ask.