Re: How to provide useful debug info

Hi,

Generally the minimum required information is /var/run/dmesg.boot and a 
stack backtrace at the crash.

See 
http://www.uk.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/kerneldebug.html, 
particularly 17.4 and 17.5

At 13:13 23/7/03, Harald Schmalzbauer wrote:
>Hi all,
>
>I have currently at least 4 scenarios when my 5.1-release crashes on
>different hardware.
>So I built a kernel (GENERIC) with debugging symbols and DDB option.
>Now I'd like to provide usefull info about the following crashes:
>
>1. booting from degraded RAID1 with HPT372 (machine crashes and I see db>)
>
>2. crash when unquirked UMASS is unpluged
>
>3. crash when /stand/sysinstall is called after some uptime
>
>4. machine hangs when LPT is removed in BIOS and ACPI enabled (no crash, I
>think I have to enter DDB with ctrl-alt-esc?)
>
>I read the FAQ (18.13.  How can I make the most of the data I see when my
>kernel panics?) but I think this isn't applicable to my crash when the
>machine crashes before rc.conf can be read. And I really don't understand
>what he's talking about :( but perhaps I can help.
>
>Please tell me if I should do something like "trace" and write it down or
>something like that.
>
>Thanks,
>
>-Harry


--
Bob Bishop		    +44 (0)118 977 4017
rb_at_gid.co.uk		fax +44 (0)118 989 4254