One or more devices are configured to use a non-native block size

[RAJOD]

Just noticed this Status warning on my newer 9.3 NAS. Its a ZFS mirror.
Is this a big deal? It seems fast enough right now. Or is this something I should correct? If so how?
=================================================
pool: ZPoolToshiba
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 0 in 5h47m with 0 errors on Wed Sep 30 19:47:50 2015
config:

NAME STATE READ WRITE CKSUM
ZPoolToshiba ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
ada0 ONLINE 0 0 0 block size: 512B configured, 4096B native
ada1 ONLINE 0 0 0 block size: 512B configured, 4096B native

errors: No known data errors
=========================================================

[ChriZathens]

During the creation of the vdev (Disks|ZFS|Pools|Virtual device|Add) you must make sure to have advanced format checked.

Capture.PNG

If you want to do it now it is not possible. You have to destroy the pool and create new vdevs with this checked.
But you are fine even if you leave it the way it is now.

[RAJOD]

Thanks for the reply,

Yea I was afraid it would require destruction of my pool. I set it up pretty fast and overlooked that checkmark. But its just for long term storage so it does not have to be a speed demon. It gets 90-100MB/sec over the lan which is good enough for me.

[ChriZathens]

You are fine, then...
No need to do anything. And, TBH, even if you did, you wouldn't see any speed difference in Lan transfers.

[RAJOD]

Apparently many have made this mistake, I guess it was a bug in the default settings of the software. The auto detect should have picked up the proper size but sometimes it fails, at least that is what I got from this thread.

https://forums.freenas.org/index.php?th ... -rc.16737/

Those people it bother and ended up destroying the pools to make it right. They really did not say what was so wrong with leaving it as is.

[Parkcomm]

here is one data point for you - http://www.tomshardware.com/reviews/adv ... 759-6.html

It shows that iops for a misaligned drive are 30% in a file server benchmark and 50% lower in a database benchmark.

However, a ZFS raid will mitigate this with the ARC.

I'd also say it's likely that the performance hit observed in the benchmark is less in newer drives.

[RAJOD]

Parkcomm,

Thanks, good article on Toms. Yes it looks like performance hit on reads or streaming is zero. I won't be doing any heavy database I/0 on it so the performance hit will be small. Maybe updating the plex database might be a tad slower. All in all not a big deal, I think I'll leave it alone.

[Parkcomm]

cool

[ChriZathens]

Honestly, the only situation where you would - probably - see some difference is while scrubbing, or resilvering.
So yeah, let it be...