"Arnaud Brand" <[email protected]> writes: > Good morning list, > > I run OpenBSD (7.8 now) on a APU4C4 board. > > I use vmd to run a small VM to host a DNS server, also running on > OpenBSD 7.8. > > This setup worked perfectly fine when I first installed it back in > March 2022 (with OpenBSD 7.0 at the time). > At some point in time ( don't remember when, sorry ) I noticed clock > drift in the VM. > My natural reflex was to setup ntp to keep the clock in sync. > This seemed to do the trick and I forgot about it. > > About a year or so ago, I noticed that there was still some clock > drift and that ntp was not able to prevent it. > I read on the web that the VM was supposed to use pvclock to stay in > sync with the host and ntp might be counterproductive. > So I removed ntp on the VM. > This didn't fix the problem though. > > I recently upgraded both the host and the VM to 7.8 and syspatched them. > The problem still occurs. > > When it occurs, I can no longer login to the VM via SSH. > > So to dig deeper, I left a console session open on the VM via a tmux > session. > This console session stays open and I can issue commands, but the > clock is frozen. > > I tried to generate a sendbug by using "sendbug -P > file". > I get an error : > cp: /var/db/acpi/*: No such file or directory > b64encode: *: No such file or directory > The file gets generated though. > > I see nothing new in the dmesg, nothing in /var/log/daemon, nothing in > /var/log/messages. > > What can I look at ? > Where can I look ? > What did I miss ? > > Thanks for your help !
Can you share a dmesg of the host machine and the output of `sysctl kern.timecounter` in the guest vm? -dv
