On 2022/05/21 11:04, Sebastien Marie wrote: > On Sat, May 21, 2022 at 04:35:25AM -0400, Johan Huldtgren wrote: > > hello, > > > > I tried to update my server from 7.0 to 7.1 tonight but after the upgrade > > booting 7.1 it just > > hangs after having mounted the filesystems. > > could you describe the filesystems you are mounting ? ffs, nfs, mfs ? showing > your /etc/fstab ? > > having the exact output at hangs time could also help to figure which step is > hanging.
you could also boot a ramdisk and modify /etc/rc to add some echos around that point to help see where it gets up to. there are a few things silently called after mounts which can hang sometimes. i am particularly wondering about ttyflags and the "com at acpi" change in 7.1. if you can get a dmesg off 7.1 (depening where the hang is, you might do it with boot -s and mounting manually) to compare with 7.0 that may give clues. > how long do you wait before considering it hangs ? (could it be possible > network timeout ?) > > is the host responding to ping ? > > > > After leaving it for a long time and it never > > proceeds past that. This is an MP system but using the SP kernel has the > > same result. > > good to know that MP and SP has the same problem. > > > I tried > > booting bsd.rd which works and installing the latest snap has the same > > result. > > good to know that -current has the same problem. > > > I then fetched > > the bsd.mp from the 7.0 directory but that fails to boot but with > > > > Process (pid 1) got signal 11 > > it is (somehow) expected. you are trying to run 7.1 userland (/sbin/init > here) > with a 7.0 kernel and it isn't a supported configuration. > > > just repeating on the screen forever. > > > > This server has been running OpenBSD for a very long time and has generally > > not had many > > problems. As I can't bring it up I transposed dmesg from looking at > > /var/run/dmesg.out > > after booting bsd.rd, any hints on what I can do to revive my server would > > be appreciated. > > Immediate revive implies reinstalling 7.0. > > > Things to try to help debugging: > > - You could try to boot on bsd.rd, go in shell, and try disabling what it is > running at boot time (/etc/rc.conf.local, /etc/rc.local). > > - adding "set -x" at top of /etc/rc could be useful too (to see the last > running > command before hang) > > - The fact that bsd.rd is fine (kernel + userland: you upgraded) might > implies > that some kernel elements in GENERIC.MP are broken. but it seems you are > reaching GENERIC.MP userland and run it successfully (/etc/rc is already > running to mount filesystems before your hang). so for now I am unsure > which > device could be disabled in boot_config(8) to test it. > > Thanks. > -- > Sebastien Marie >
