SuperMicro X8DTN+ and SAS-826EL1 Compatible?

[Link Up]

zero or more pools/volumes, each composed of one or more vdevs, each composed of one or more HDs.

You can add pools at any time, and you can add vdevs to pools at any time.

You can't add/remove HDs to/from a RaidZ[123] vdev.

And you can't remove vdevs from a pool.

Is that to say that if for instance I have a vdev made up of 4 2TB drives, that I can't upgrade the drives to a bigger size? I could presumably replace a bad 2TB with another drive, but not expand the vdev, either by enlarging the drives or adding more drive to the vdev?

That would seem to make a person not want to have more than one vdev per pool to avoid inherent limitations? That would then mean you couldn't have one large volume?

Thanks!

 

[danb35]

Is that to say that if for instance I have a vdev made up of 4 2TB drives, that I can't upgrade the drives to a bigger size?

Of course you can, and this is even described in the manual.

 

[Stux]

Is that to say that if for instance I have a vdev made up of 4 2TB drives, that I can't upgrade the drives to a bigger size? I could presumably replace a bad 2TB with another drive, but not expand the vdev, either by enlarging the drives or adding more drive to the vdev?

That would seem to make a person not want to have more than one vdev per pool to avoid inherent limitations? That would then mean you couldn't have one large volume?

Thanks!

You can replace HDs in a vdev. If all the HDs in a vdev are replaced with larger HDs, then the vdev will expand (if autoexpand is true)

 

[Link Up]

You can replace HDs in a vdev. If all the HDs in a vdev are replaced with larger HDs, then the vdev will expand (if autoexpand is true)

I was trying to reconcile that with your previous statement:
"You can't add/remove HDs to/from a RaidZ[123] vdev"

It seems you can remove a failed drive, and you can add a replacement drive back, so on the surface, what you stated has to be qualified. It still seems based on what both of you have said, it makes sense to have a pool per vdev so if you wish to remove drives or make other? changes, nothing goes to heck.

Thanks to both of you!

 

[danb35]

"You can't add/remove HDs to/from a RaidZ[123] vdev"

To state that more precisely, once a RAIDZn vdev has been created with x disks, it can't be changed to a (healthy) RAIDZn vdev with other than x disks. Disks can be removed, which will degrade the vdev, and disks can be added to replace disks thus removed. And if you replace a failing disk before it completely dies, the vdev can temporarily have more than x disks. But the healthy condition of that vdev will always be x disks, never x - y or x + y.

 

[Stux]

I was trying to reconcile that with your previous statement:
"You can't add/remove HDs to/from a RaidZ[123] vdev"

It seems you can remove a failed drive, and you can add a replacement drive back, so on the surface, what you stated has to be qualified. It still seems based on what both of you have said, it makes sense to have a pool per vdev so if you wish to remove drives or make other? changes, nothing goes to heck.

Thanks to both of you!

Thing is, when you remove a failed drive, the drive is still there, but offlined, until it is replaced.

Anyway, the point of my post was to inform you that you can't easily increase a 6-way Raidz2 to a 7-way or greater Raidz2. Etc

 

[Link Up]

Thanks again to both of you. I am spoiled by my ReadyNAS which allows you to add drives until the bays are full and any drive can be pulled out and replaced with a same size or larger drive.