ZFS non-native block size after Update

[psycho160]

Hello,

i have a zfs pool (ashift=9) and two different disks, one Seagate 4K (512e) and a Toshiba 4K (512e)

My zfs setup was running about 5 months with 6.3 and after the latest proxmox update yesterday my check_mk put a warning about my zfs pool.
So i looked into pve and zfs status says:

root@pve01:~# zpool status
  pool: tank
state: ONLINE
status: One or more devices are configured to use a non-native block size.
        Expect reduced performance.
action: Replace affected devices with devices that support the
        configured block size, or migrate data to a properly configured
        pool.
  scan: scrub repaired 0B in 04:05:23 with 0 errors on Sun Mar 14 04:29:24 2021
config:

        NAME                                 STATE     READ WRITE CKSUM
        tank                                 ONLINE       0     0     0
          mirror-0                           ONLINE       0     0     0
            sda                              ONLINE       0     0     0
            ata-ST4000DM006-2G5107_ZC17WAL3  ONLINE       0     0     0  block size: 512B configured, 4096B native

errors: No known data errors

then i looked at:

root@pve01:~# cat /sys/class/block/sdd/queue/physical_block_size
4096
root@pve01:~# cat /sys/class/block/sdd/queue/logical_block_size
512
root@pve01:~# cat /sys/class/block/sdc/queue/physical_block_size
512
root@pve01:~# cat /sys/class/block/sdc/queue/logical_block_size
512

SDD is the Seagate disk, with a pysical block size of 4096...

What can i do to get rid of this error/warning? And what does "Expect reduced performance." mean? I did only the proxmox update, nothing related to the zfs....

 

[leesteken]

The "reduces performance" is not new, only the message is new from the updated ZFS version that comes with Proxmox updates.
To fix this warning you could stop using that SSD for pools with ashift=9 and only use it for pools with ashift=12.
EDIT: Maybe remove that SSD from tank and create a new pool on it with ashift=12, copy everything over, destroy tank and add sdc as a mirror to the new pool?

 

[psycho160]

The "reduces performance" is not new, only the message is new from the updated ZFS version that comes with Proxmox updates.
To fix this warning you could stop using that SSD for pools with ashift=9 and only use it for pools with ashift=12.
EDIT: Maybe remove that SSD from tank and create a new pool on it with ashift=12, copy everything over, destroy tank and add sdc as a mirror to the new pool?

only for correction, my disk is not an SSD as you said, it´s an HDD

I´m afraid if i rebuild the tank, that then the other disk would alert that i use a ashift with 4K and the disk reports only 512B..

 

[leesteken]

only for correction, my disk is not an SSD as you said, it´s an HDD

I´m afraid if i rebuild the tank, that then the other disk would alert that i use a ashift with 4K and the disk reports only 512B..

Sorry, I must have read /dev/sdd as SSD. No need to worry, I have that situation and it does not complain about it.
Writing 8 sequential 512-bytes sectors for each 4096 bytes also performs better than reading 4096, changing only 512 bytes and then writing 4096 each time.

 

[psycho160]

thx, for the info.
i found an amazon product comment where this error is described with especially toshiba disks. maybe a firmware issue
Amazon comment

i will try to rebuild my tank