https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=288345

--- Comment #18 from Craig Leres <[email protected]> ---
I've had many more hangs but today was the first time I was able to get on the
(ipmi) console and try to get a crash dump. So far I'm still unsuccessful.

I was getting usage messages if I added -Z to savecore so I left it out. I
tried this:

    # savecore -kv
    unable to open bounds file, using 0
    checking for kernel dump on device /dev/mirror/gm0p2
    mediasize = 34359737856 bytes
    sectorsize = 512 bytes
    magic mismatch on last dump header on /dev/mirror/gm0p2
    savecore 43 - - no dumps found

"dumpon -l" showed /dev/null but savecore was still able to locate mirror/gm0p2
as the dump device (which is what I see when the system is booted multiuser).

I also see that when the system is running, I've configured /var/crash as a zfs
partition but it's not mounted when /etc/rc.d/savecore runs. I guess having
dedicated zfs crash partition isn't really possible.

Is something initializing swap when I boot into single user after "sysctl
debug.kdb.panic=1"?

This is a pretty annoying bug that I only see on this one system and I'm not
sure what to try next time it occurs.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to