Added subject.
Thank you. It escaped me for a moment that I was composing an email and
thought "Synopsis ✔️".
ddb> show panic
*cpu0: kernel diagnostic assertion "amap->am_buckets[bucket] == NULL"
failed: f
ile "/usr/src/sys/uvm/uvm_amap.c", line 147
I have no clear explanation and have seen various panics related to
amap
corruption. This is just one of them.
If possible, are you able to try running snapshots? Otherwise, when the
next release is cut, the redesign will be in vmm(4).
TBH I do not understand what's this about, but I can try to run
snapshots to help.
I could only find a Stackexchange answer [1] on how to switch, but it
does read more complex than I'd expect from OpenBSD. Then, I'm actually
on resflash and I'd have to figure out how to track snapshots long-term
there.
A bigger worry that came up with this is about using OpenBSD as a
hypervisor. As far as what I read, people say they only ever lost one
file or another in FFS2, but im case of virtualization this one file is
an entire disk.
So far I've lost two due to this bug: The first luckily early in setting
up the VM and the second was a mount to a subdir in /var/lib which could
be re-created with some three commands.
Where I'm getting at: Even once this panic is solved an unexpected power
outage will probably cause the very same type of corruption.
Nevertheless I'm stubborn enough to give it a chance. I will just have
to do a lot of backups :)
[1]
https://unix.stackexchange.com/questions/411202/openbsd-how-to-follow-current
