tsoome_, I had an issue with piadm on SmartOS few weeks ago. I decided to remove piadm from the pool and boot from USB flash. I had latest and some previous snapshots in piadm and then removed it from booting from hard drive.

Problem arised when USB flash was not easily bootable (SD card in USB reader) and won't boot every time on hardware.

So I decided to bring back booting from the pool again, after booting from the flash.

But, installing back and booting from the pool didn't work, because same-named snapshot were still available on disk, since installation on the pool needs to set the snapshot

So I needed to activate older snapshot and then I could remove latest snapshot and only after that I could bring back pool to bootable state again with the latest snapshot

So maybe such situation when there is already latest snapshot on the pool and one can't get pool bootable again, could be avoided ? (so that someone less savvy won't end up with inability to boot form pool)

I'm trying to create an LX zone from the latest Rocky 9 image and am getting these failures during provisioning which (I think?) are causing a provisioning timeout: gist.github.com/natefoo/b8b4bf368beda3aad75402b5cc47dfd0

Is this an image problem?

(latest Rocky 9 LX image on images.smartos.org, not one I've built myself)

Well that's not the cause - I created a centos-7 zone, it has the same errors but boots fine.

natefoo: A provision is considered successful when /var/svc/provisioning is renamed to /var/svc/provision_success. If something is delaying or preventing that, you'll end up with provision failures.

You can "cheat" this by moving it yourself early in the user-script.