Ashift 0 and other oddities with pool created in USB enclosures

Running a pool with a pair of 4-wide Z1 Vdevs on a Mini-R. The pool was previously built on a Mini-E, which didn’t have space for 8 disks, so USB enclosures were used. Now, none of the drives show as being in the pool in the enclosure view (grey) but pool itself works fine, and newly replaced disks show in the view. I previously made a topic about this in the old forum, i mention here in case it has a related cause.

Now, i have decided to check that the ashift is sane, but when running zpool get ashift, i get back 0, likewise for the bootpool, which is on an NVMe. This is with the default install as shipped from the factory.

Any ideas on why i get this bad ashift value?

A value of 0 is probably “auto”.

You could use zdb to grab the real values, but zdb is “broken” on Core.

EDIT: You can try these two commands.

For all pools (except the boot-pool):

zdb -U /data/zfs/zpool.cache | grep "ashift\|\ name:"

For the boot-pool:

zdb -e -C boot-pool | grep "ashift\|\ name:"
3 Likes

That was very helpful, thank you. :smiley:

Luckily, the pool I was concerned about is ashift 12, but strangely the boot-pool is 9. I’ll probably leave that alone, but still strange.

ashift 9 for the boot-pool is probably a remnant from the early days of FreeNAS.

iX could probably default to 12 for the boot-pool, but maybe they want to leave it open for USB sticks?