• Lloyd [2025-01-02 02:03]:
> I've been doing some testing against Wireguard and was able to get the system
> in a state where it was unable to reboot via the shutdown command. Unfortunately,
> I've been unable to reproduce the issue again so far. This particular system was
> running an unmodified GENERIC kernel from -current:
>
> OpenBSD 7.6-current (GENERIC) #470: Thu Dec 12 09:20:35 MST 2024
> deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC
>
> I had the system sending small amounts of traffic over the tunnel overnight. I
> had executed an 'ifconfig wg0 down' then attempted an 'sh /etc/netstart wg0'.
> Not sure if this had succeeded, I attempted to reboot the box, both using 'reboot'
> and 'shutdown -r', neither of which worked. It would get as far as "System
> shutdown time has arrived" and return you to the shell prompt. A few other
> peculiar things I noticed:
>
> - ifconfig would hang during the display of the wg0 interface, hanging after the
> line "index A priority B llprio C" and a ^C would not terminate it. ifconfig
> had to be manually killed
> - the /usr/local/bin/wg process was unable to be killed, via kill -9 or kill -15
> - a few attempts at running reboot and shutdown showed multiples of those
> processes (and rc) hanging in the process list.
>
> Can't remember what exactly I did that finally broke the hang and allowed the
> reboot to proceed, but it involved tickling the wg0 interface. I don't have
> enough concrete data for a bug report but seeing if anyone has any ideas.
FWIW I've seen wg destroy bring system to state where reboot didn't
work: https://marc.info/?l=openbsd-tech&m=169642385205901&w=2 with my
steps to reproduce in
https://marc.info/?l=openbsd-tech&m=169644634001172&w=2
One of the responses mentioned this go back to at least 7.2 in October 2022:
https://marc.info/?l=openbsd-tech&m=169642819009012&w=2
And bug report here:
https://marc.info/?l=openbsd-bugs&m=166724872311640&w=2
> Regards
> Lloyd
>
No comments:
Post a Comment