It happens all the time. A kernel version upgrade pulls ZFS, and then it hangs on reboot, because ZFS will wait forever. The only solution at this point is to reboot in emergency mode and remove ZFS. Is there any way to preclude this behavior? The server does not have any ZFS pools. This is a bug.
Possibly a workaround for you: you can blacklist kernel-modules. This should also work for
zfs and spl. You should be able to find a "how to" in the PCI-passthrough context.Disclaimer: never tried that; I've never had such problems...
Could this be one of the pitfalls (where not all dependencies are updated) of running apt-get upgrade instead of the recommended apt-get dist-upgrade?
Unfortunately not as we update all nodes for years by alias without issues (3 without zfs while 2 with zfs)
alias update='apt update && apt upgrade -y && apt autoremove -y'
but there was obviously something going bad in the past of falves1 installation in the past to get this happen result.
Maybe there was an experimental additional zfs pool which wasn't completely removed from a started systemd import unit for ? Just as an idea ...
alias update='apt update && apt upgrade -y && apt autoremove -y'
but there was obviously something going bad in the past of falves1 installation in the past to get this happen result.
Maybe there was an experimental additional zfs pool which wasn't completely removed from a started systemd import unit for ? Just as an idea ...