Hi Slowservers, There's been a thread some months ago on misc@. I solved it by using this config in ntpd.conf in the guest. sensor vmmci0 refid LOCL stratum 8 trusted Have a good one ! Arnaud On 29/08/2026 06:39:34, [email protected] wrote:
Hi bugs@, Something weird is going on with one of my 7.9-on-7.9 VMM guests. I migrated one of my OpenBSD guests from an Intel 7.8 host to a AMD 7.9 host with scp on the disk image. Guest is running nsd, Caddy, httpd, and Prosody. After a few days running, I started having some issues with it. Caddy is responsive. SSH never finishes the handshake, but readily returns the header. ICMPv6 echo is replied to. Guest was stable for a few days. And on the prior host, stable for several months. I'm able to login to the guest via the serial console and this is what I found: potato# w 9:15PM up 3 days, 21:31, 1 user, load averages: 0.32, 0.09, 0.02 USER TTY FROM LOGIN@ IDLE WHAT root 00 - 9:15PM 0 w potato# w 9:15PM up 3 days, 21:31, 1 user, load averages: 0.32, 0.09, 0.02 USER TTY FROM LOGIN@ IDLE WHAT root 00 - 9:15PM 0 w potato# w 9:15PM up 3 days, 21:31, 1 user, load averages: 0.32, 0.09, 0.02 USER TTY FROM LOGIN@ IDLE WHAT root 00 - 9:15PM 0 w potato# w 9:15PM up 3 days, 21:31, 1 user, load averages: 0.32, 0.09, 0.02 USER TTY FROM LOGIN@ IDLE WHAT root 00 - 9:15PM 0 w potato# date Fri Aug 28 21:15:22 UTC 2026 potato# date Fri Aug 28 21:15:20 UTC 2026 potato# date Fri Aug 28 21:15:21 UTC 2026 potato# date Fri Aug 28 21:15:21 UTC 2026 potato# date Fri Aug 28 21:15:22 UTC 2026 potato# date Fri Aug 28 21:15:23 UTC 2026 potato# date Fri Aug 28 21:15:24 UTC 2026 potato# date Fri Aug 28 21:15:20 UTC 2026 potato# date Fri Aug 28 21:15:21 UTC 2026 potato# date Fri Aug 28 21:15:22 UTC 2026 potato# date Fri Aug 28 21:15:22 UTC 2026 potato# date Fri Aug 28 21:15:22 UTC 2026 potato# date Fri Aug 28 21:15:23 UTC 2026 potato# date Fri Aug 28 21:15:23 UTC 2026 potato# date Fri Aug 28 21:15:20 UTC 2026 potato# date There is nothing interesting in the guest's dmesg. The load average never changes and the date is stuck looping. I killed ntpd and it's still doing it. I had a customer guest that I migrated from the same 7.8 Intel host to a 7.9 AMD host. That customer also experienced something weird, maybe the same thing, that I haven't been able to reproduce yet. I suspect it is the same issue for different reasons. I've had rock solid OpenBSD guests on my Intel hosts, and rock solid Debian guests on the AMD hosts. But it seems that something is strange on the AMD host. The time on the hosts is normal. My AMD hosts have 4162 EE Opterons, if it matters. I am quite confused on what is going on here. Thank you! -Slow Servers PS: I am not on this list, so please CC me on replies.
