FN 9.3 "block size: 512B configured, 4096B native"

[MtK]

Hey,
I just replaced a faulty disk with a brand new one in my RaidZ2 and get this:

Code:

# zpool status
  pool: Storage
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: resilvered 219G in 3h29m with 0 errors on Tue Feb 17 19:33:01 2015
config:

NAME                                            STATE     READ WRITE CKSUM
Storage                                         ONLINE       0     0     0
  raidz2-0                                      ONLINE       0     0     0
    gptid/64d80576-5a7d-11e3-bce5-002590d91105  ONLINE       0     0     0
    gptid/651b51ee-5a7d-11e3-bce5-002590d91105  ONLINE       0     0     0
    gptid/655ff841-5a7d-11e3-bce5-002590d91105  ONLINE       0     0     0
    gptid/65a2fa41-5a7d-11e3-bce5-002590d91105  ONLINE       0     0     0
    gptid/65dfc343-5a7d-11e3-bce5-002590d91105  ONLINE       0     0     0
    gptid/1cacb72e-b6b6-11e4-adc4-002590d91105  ONLINE       0     0     0  block size: 512B configured, 4096B native

The pool already resilvered...

Do I have a way of solving this?

 

[MtK]

Just to be clear, the disk replace was done through the GUI on the latest FN 9.3-STABLE.

 

[dlavigne]

That's interesting... Might be worth making a bug report at bugs.freenas.org. If you do, post the issue number here.

 

[cyberjock]

This is expected. Your zpool uses 512 bytes/sector. Newer disks use 4k sectors. So you aren't matching up sector sizes. The solution? Just accept the lower performance or create a new pool that is all 4k sector size disk with an ashift value of 12.

 

[allegiance]

I had the same issue and was looking for a response.

So, if I were to create a new pool that uses all newer disks that have a 4k sector size, will the GUI build the pool with the matching 4k sector size, or would I need to specify it somehow?

 

[cyberjock]

FreeNAS after 9.2.1.0 (I think that version anyway) will *always* create pools with an ashift of 12. This is because there's no reason to assume you want to use an ashift of 9 when you aren't going to buy disks with 512 bytes/sector anymore. So time to get with the rest of us and move on to 4k.

For those of you with the ashift of 9, you should try to find a way to migrate to a pool with an ashift of 12 (you can't "convert" a pool to a different ashift value).

The reality is that the performance penalty for users that mix and match this stuff, in a home setting, is going to be indistinguishable from network latency. But if you plan to run a dozen high-intensity VMs on your machine, then you might have a problem.

What would I do if I was MtK? I'd leave the pool exactly how it is right now. Just keep in mind that someday you'll want a bigger pool, and it'll be time to build a new pool with new disks and copy your data over so you can enjoy all that 4k goodness.

 

[MtK]

What would I do if I was MtK? I'd leave the pool exactly how it is right now. Just keep in mind that someday you'll want a bigger pool, and it'll be time to build a new pool with new disks and copy your data over so you can enjoy all that 4k goodness.

In my case I have a very small (home) pool which at some point will be replaced with new 4Tb drives.
So yes, that's the plan, wait for that moment, move everything to a backup pool, rebuilt the pool with new drives and put data back...


thanks!