Maximum Drives

[burnmai]

Hi,

Quote from the User Manual:
"
When determining how many disks to use in a RAIDZ, the following configurations provide optimal
performance. Array sizes beyond 12 disks are not recommended.
Start a RAIDZ1 at at 3, 5, or 9 disks.
Start a RAIDZ2 at 4, 6, or 10 disks.
Start a RAIDZ3 at 5, 7, or 11 disks
"
What do they refer to with "Array".
Do they mean drives in a pool?
Do they mean drives in a physical Vdev
Do they mean physical Vdevs in a logical Vdev?

Best regards.

 

[cyberjock]

Do they mean drives in a pool?

No, drives in a vdev. Pools aren't in RAIDZ1, RAIDZ2 or RAIDZ3.. so trying to say a pool that is RAIDZ2 isn't entirely accurate, but most people will interpret the user to mean a pool that is a single vdev that is RAIDZ2.

Do they mean drives in a physical Vdev
Do they mean physical Vdevs in a logical Vdev?

No clue where you got the idea there's such a thing as "physical vdevs" versus "logical vdevs" since there's only one kind of vdev... "a vdev".

 

[burnmai]

I got stuck on the "Array" part, because its not defined enough

ZFS has two types of vdevs: logical and physical. So, from the ZFS on-disk specification, we know that a physical vdev is a writeable media device, and a logical vdev is a grouping of physical vdevs.

That is the difference :D

 

[cyberjock]

Yes, but they don't actually matter for the scope of your question. If you want to start talking about disk formatting that's far deeper than what you need to be concerned with.

I do see your confusion about "array". I'll fix up the manual when I get a chance. :)

Edit: I see what is going on, it's a hodge-podge because the manual still covers things for UFS and ZFS. I will make sure that when 9.3 hits the documentation is a little easier/more accurate with its discussion. UFS is gone so there's no reason to discuss hardware RAID at all except to provide the comparisons between hardware RAID and ZFS so people can grasp the fundamentals more easily.

 

[burnmai]

No clue where you got the idea there's such a thing as "physical vdevs" versus "logical vdevs" since there's only one kind of vdev... "a vdev".

Yes, but they don't actually matter for the scope of your question. If you want to start talking about disk formatting that's far deeper than what you need to be concerned with.

That was due to the fact that i was looking for an explanation for the "Array" I thought i missed something and Basic went trough 50% of Suns documents about ZFS and the Manual that Freenas offers... at least i found the solution.

But this raises the question, what does happen if i (theoretically speaking) create a Vdev of 20 disks?

You stated: (in a different thread)

No, vdevs should never be bigger than 11 disks... ever. I went with an 18 drive vdev and there's a laundry list of reasons I won't go into detail with here as to why you shouldn't do it.. You need to go to 2 vdevs of RAIDZ2. 6 drive vdevs is ideal, or if you had 1 more disk you could do 8 disk vdevs(which is non-optimal).

What did you stumble upon?

 

[cyberjock]

You get so wide that you end ip IOPs bottlenecked pretty badly. Performance degrades and it just sucks. Resilvering the pool when a disk fails makes the pool almost unavailable for "high performance" tasks.

Initially it seems like all is well, but once you load it with data it's very disappointing. I won't lie.. initial benchmarks were good. Then things started going downhill fast.

 

[dlavigne]

I've reworded that section for 9.3 to not use the word array.