Issues with openbsd 7.8 arm64

It looks like your system clock is off. Can you try syncing the time via NTP?

This usually happens when the date and time on the VPS are wrong. Set the correct date/time and you should get rid of that warning i guess.

You should be able to set the time and date.

If i remember it correctly the installer expects input right there. You can type the time in a format like: 2025-12-15 14:30:00 and hit Enter.

This corrects the clock so the install can continue.

(Someone correct me if i am wrong please)

Nevermind i guess i was wrong here.

When (or if) you reach the installer prompt, choose (S)hell

In the shell, manually set the date using the BSD date command format: date yymmddhhmm[.ss]

Example: date 2512151430

If i am wrong again i am sorry, i didnt use OpenBSD for a long time

But do you really need to fix it at boot time given it's a warning and not a fatal error? Once the installation is finished and you reboot, there should be an NTP daemon which takes care of this issue. (And even during the installation process, you normally would be able to drop to a shell and invoke the "date" command.)

Hi. Had the same problem.

You need to do a little dance with a kernel config on cd image, see this thread: https://marc.info/?t=170233621600002&r=1&w=2

long story short, kernel decides to ignore vga and use serial console only, and netcup gives us vnc-to-vga access only. you probably can try and set output/input/error device in uefi setup, but i'm not sure if it works - so the easier way is to edit bsd.rd, disable `pluart` device, and put it back on cd., see second to-last message in thread. aaand you need an arm64 openbsd box to do it - so boot a qemu vm, ftp https://cdn.openbsd.org/.../cd78.iso, then vnconfig vnd0 cd78.iso, mount_cd9660 /dev/vnd0c /mnt, cp /mnt/7.8/arm64/bsd.rd /tmp/cd.bsd.rd, then you need to use config like this:

# config -e -f /tmp/cd.bsd.rd
OpenBSD 7.8 (RAMDISK) #38: Sun Oct 12 18:49:37 MDT 2025
deraadt@arm64.openbsd.org:/usr/src/sys/arch/arm64/compile/RAMDISK
Enter 'help' for information
ukc> find pluart
142 pluart* at rpone*|apldc*|agintc*|ampintc*|qcdwusb*|imxsrc*|imxdwusb*|mvmdio*|rktcphy*|rkpinctrl*|rkgrf*|rkdwusb*|hidwusb*|amldwusb*|syscon*|sxisyscon*|simplebus*|mainbus0 early 0 flags 0x0
294 pluart* at acpi0 addr -1 flags 0x0
ukc> disable 142
142 pluart* disabled
ukc> disable 294
294 pluart* disabled
ukc> quit
Saving modified kernel.

now - create directory structure for cd and run mkhybrid to put it back together:

# mkdir newcd
# mkdir -p newcd/etc newcd/7.8/arm64
# mv /tmp/cd.bsd.rd newcd/7.8/arm64/bsd.rd
# cp /mnt/etc/boot.conf newcd/etc
# cp /mnt/7.8/arm64/cdbr newcd/7.8/arm64/cdbr
# mkhybrid -a -R -T -L -l -d -D -N -o newcd78.iso -b 7.8/arm64/cdbr -c 7.8/arm64/boot.catalog newcd/

(i took that mkhybrid call straight from Makefile, i have no idea what first ~10 options do)

et voila! you should have newcd78.iso that's should be bootable under netcup's KVM setup.

P.S. It would be really great if someone figures out whose to blame for that behaviour and fixes it. Or maybe netcup will do same stuff i described and update their provided iso file.

P.P.S. Better yet - why not provide us with serial console over ssh or something instead of VNC? or maybe even bare ipxe scripts... mmm...

hope this helps!

got distracted and forgot to share actual iso file i made:

https://pub.pzskc383.net/openbsd-78-arm64-netcup.iso
SHA256 (openbsd-78-arm64-netcup.iso) = 5ba3a0438195d13e5466ef9013033066045caaf3a35d123027d75fa3cdf894c5

tested on my server - boots to console fine.