If you’ve been experimenting with ZFS, as we have, you’ve been struck by some of its limitations. One of the most common complaints is that it’s not flexible in terms of sizing – you have to add a whole new “vdev” (analogous to a RAID set) in order to expand, rather than simply adding new disks one at a time.
But there is one other way to expand a zpool – by iteratively swapping out every disk in the pool with a larger one and letting ZFS heal the pool. This is the technique demonstrated by Harry D in a two-part blog post:
You still can’t shrink a zpool. Let’s hope that comes along soon, along with more flexibility in general. Otherwise, ZFS will be much less interesting in the desktop/workstation/small server space!