News for hackers. Stuff that matters.
Re:ZFS lands in Debian contrib by wtallis on Saturday May 14, 2016 @08:10AM (#11695745)
You're describing operations that can be done on a pool, while conveniently ignoring that they can not be done to a single RAID array (vdev), which is the more useful comparison against btrfs.

Yes, you can add drives of arbitrary size to a ZFS pool, but when put into the same array, they all get treated as the size of the smallest one. Yes, you can expand a ZFS pool but only by adding a new array or by incrementally upgrading all the drives in the new array. Yes, you can remove drives, but only for replacement purposes; you can't do an in-place rebuild to make the reduced number of drives the new normal instead of degraded mode.

Re:ZFS lands in Debian contrib by simoncion on Saturday May 14, 2016 @08:21AM (#11695777)
I know nothing about ZFS, so sorry if these are dumb questions.

Can you have an array that contains a single disk? And if you can, then

* Can you add as many single drive arrays to a pool as you like?

* Can you migrate data off of a given single-drive array on to other single-drive arrays in the pool and then permanently remove the array from the pool?

Re:ZFS lands in Debian contrib by lmz on Saturday May 14, 2016 @08:27AM (#11695794)
Yes, yes, no. Plus a single disk array has no redundancy.
Re:ZFS lands in Debian contrib by X86BSD on Sunday May 15, 2016 @09:15AM (#11700783)
Minor nit: depending on your definition of redundancy, you may set copies=n, on a single vdev pool. You won't get hardware redundancy but will get data redundancy.