Date: Wed, 18 Aug 2004 07:31:01 +0200 From: =?ISO-8859-1?Q?Eirik_=D8verby?= <ltning@anduin.net> To: Julian Elischer <julian@elischer.org> Cc: stable@freebsd.org Subject: Re: How to find the cause of a hang Message-ID: <CA6BB5FA-F0D7-11D8-AE9C-000D9335BCEC@anduin.net> In-Reply-To: <411E7B28.7050505@elischer.org> References: <26ECE35D-EDDB-11D8-945B-000D9335BCEC@anduin.net> <411E7B28.7050505@elischer.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 14. Aug 2004, at 22:50, Julian Elischer wrote: > Eirik Øverby wrote: >> Hi all, >> I'm currently experiencing frequent (about once per week) hangs of a >> server that is about 1500 kilometers away from me. I have a serial >> cable on the box, and using minicom on the neighbor box I am now in >> the kernel debugger - but I'm at a complete loss as to what to do to >> figure out what is, in fact, wrong. >> Calling panic or boot doesn't work - it just stops at "syncing >> disks..." and never actually reboots. I suspect something fishy going >> on with disk I/O, but I can't be certain of that. >> The box responds to ping - until I call panic or boot - but no other >> services are working. > > try capture a stack trace "tr" Looks like the box is "idle" > if you have KTR enabled do "show ktr" This is 4.x... As my follow-up msg on current@ indicated (yes i posted to the wrong list initially ;) > do "ps" An insane number of cron processes.. Is it trying to run scheduled jobs and fails because of the hang? > do show pcpu > show witness > show locks No workie on 4.x... > if you have a dump device define.. > call doadump No dump device here (if disk is the problem, it would be of no use I guess), and i get undefined symbol anyway. > then to reboot.. > "call cpu_reset" Now that one is handy. ;) > The dump will appear after the next boot in /var/crash > if it's not big enough for a complete ram dump, symlink it to > somewhere where there is enough room. (See above) > when you have all that.. let us know :-) I have what I have. See http://anduin.net/~ltning/debug.cap (it is too long to include in a mail... Got any clues for me? Thanks, /Eirik > > >> What can I do? I'm now at the db> prompt ... Help :) >> /Eirik >> _______________________________________________ >> freebsd-current@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-current >> To unsubscribe, send any mail to >> "freebsd-current-unsubscribe@freebsd.org" > > > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CA6BB5FA-F0D7-11D8-AE9C-000D9335BCEC>