nixosTests.zfs.unstable is failing for some reason, it seems like boot.zfs.enableUnstable which the test uses has no effect? O.o

@ofborg test zfs.unstable

@Atemu The test passes for me locally:

nixpkgs on  pr-108944 λ nix-build ./nixos/release.nix -A tests.zfs.unstable.x86_64-linux
...
machine # [   19.117462] dhcpcd[990]: eth0: adding default route via fe80::2
machine # [   19.362872] zed[991]: eid=1 class=pool_create pool='rpool'
machine # [   20.059297] zed[1044]: eid=2 class=history_event pool='rpool'
machine # [   20.085840] zed[664]: Missed 1 events
(1.78 seconds)
machine: must succeed: zfs create -o mountpoint=legacy rpool/root
(0.26 seconds)
machine: must succeed: mount -t zfs rpool/root /tmp/mnt
(0.07 seconds)
machine: must succeed: udevadm settle
(0.08 seconds)
machine: must succeed: umount /tmp/mnt
machine # [   20.600575] systemd[1]: tmp-mnt.mount: Succeeded.
(0.08 seconds)
machine: must succeed: zpool destroy rpool
machine # [   20.726437] systemd[1]: tmp-mnt-rpool.mount: Succeeded.
(0.20 seconds)
machine: must succeed: udevadm settle
(0.07 seconds)
(23.30 seconds)
test script finished in 23.35s
cleaning up
killing machine (pid 6)
(0.00 seconds)
/nix/store/gysai8sqhlyx07wy0syd6va1y7a4jbbh-vm-test-run-zfs-unstable

nixpkgs on  pr-108944 took 16m 39s λ echo $?
0

Did you merge your kernel bump locally too? The zfsStable it (erroneously) uses is still compatible with the current _latest = 5.9.

Does the test use 0.8.6 or 2.0.1 on your machine? (grep the log, ZFS prints its version)

@Atemu I misunderstood. I thought this PR alone would break the test. Let me see...

I see what you mean. It seems there is an issue with how the test is evaluated. Regardless of enableUnstable, Nix tries to pull in zfsStable

@Atemu Can you post the error message? I actually don't see why it should fail here. The userspace tool in the system environment should be the one from unstable as well.

From what I can tell (I am running zfsUnstable for my root), the user space tools are indeed 2.0.x. However, part of the issue is that the stable tools are still evaluated, which causes errors from the assertions when nixos-rebuild boot.

Reminder: Please review!

This Pull Request is awaiting review. If you are the assigned reviewer, please have a look. Try to find another reviewer if necessary. If you can't, please say so. If the status is not accurate, please change it. If nothing happens, this PR will be should be put back in the needs_reviewer queue in one day.

Note: This feature is currently broken. The bot will not actually change the status. If you see this message multiple times, please request a status change manually.

Hm, now it's duplicated across release-20.09 (this PR) and staging-20.09 (c021683). But I guess that won't be a problem, since they're the same thing.
EDIT: nvm, I missed some bits on the staging backport…