After updating from fedora-38 to fedora-39, I have some AppVMs that I am unable to boot because qubes-relabel-rw takes more than 60s.
The VMs have large /rw/bind-dirs/var/lib/docker.
I can clean my docker state before making the template switch, but it would be nice to avoid having to recreate the state.
It would be good with a more user friendly process for this, both for explaining the problem and the solution.
Also, if this was a rarely used appvm, then I might not have noticed that the problem started due to a template vm upgrade.
I also might not have the old template vm present on my system to switch back to for cleaning up the bind dir.
So it would be nice with another solution for this.
The simplest to implement is probably to remove the time out, and have the user kill the vm if it takes too long.
But that might be problematic in the case of service vms that start at boot? If so, the time out could be conditional.
Rune
Thanks for suggesting this - I'll have a think, and see what might be
best suggestion for most users.
unman