ZFS zombie pool haunting me, please send help (ZFS bugs?)

[Ellie <el%horse64.org@localhost>]

Dear NetBSD list,

A haunted ZFS pool came back to end me just in time for Halloween, on my 
computer #1:

https://horse64.org/temp/100_7083.JPG

zpool import -a doesn't list it either.

Why don't I just use -f to override it...?

Simple, another computer #2 (since this is a USB disk) also has 
zombified dreams of owning this pool, even though I used zpool export 
there to disown it. At least from all I could find, an export should 
disown t, or is that incorrect?

And I already tried recreating the pool -f on computer #1 before, and 
simply booted in both systems #1 and #2 (remember, #2 has disowned a 
*previous* version of the pool, so should have doubly no stakes) and..

...it was corrupted on system #1. "no dataset available"

So yeah, I'm stuck. There seem to be multiple potential NetBSD ZFS bugs 
here, if I'm not misunderstanding the commands:

1. I feel like either ZFS should treat a partition as dead unrelated 
data and not touch nor ever mention it as a past pool, OR zpool import 
-a should list it in whatever degraded state. This is clearly not the 
case. That feels like a bug, perhaps a design bug.

2. When a zpool is created using a /dev/dkX numbered wedge, then zpool 
export'ed, then seen on a new system with a different /dev/dkX wedge 
(but same partition, can happen e.g. if it's on USB and devices end up 
enumerated differently), then the new system sees nothing with zpool 
import -a either. This also seems like a pretty hard to work around bug.

3. When a zpool was no longer listed and a new one is created on top 
with -f, none of the systems with knowledge of the no longer listed old 
pool should corrupt the data of the now new pool. This also doesn't seem 
to be the case and seems like a pretty frustrating bug too.

So how do I get rid of this zombie pool on all devices to recreate it 
free of corruption?

Regards,

Ellie