ZFS on Linux - different disk sizes in pool - want bad block healing
 

I've been testing a zfs pool using 3 drives I had left over from a computer that I got rid of. Two of them are 160G and the other is 400G.
Is it possible to raidz or JBOD them together using the whole disks in such a way that I will still have the ability for zfs to find and correct file errors (healing).

I've been searching for and reading all of the ZFS documentation I can find, but so far nothing has addressed this specific case.

The simplest approach would be to create a raidz with the three disks, that would roughly make a 320G pool with 240 GB wasted.

You might also create a 320 G stripe with the smallest disks, and mirror them with the large one, that time 80G are wasted and performance is better.

If you really don't care about performance, you might create two partitions on the large disk, one 160G and one 240G, create a 320G raidz pool with 3x160G and use ditto blocks on the 240G to create a partially redundant 120G pool, total: 440G. Note that the 120G pool won't survive the 400G full disk failure.

Similarly, you might recover 40G from the previously wasted 80G (suggestion #2), total: 360G