[no subject]

[Ignatios Souvatzis <ignatios%cs.uni-bonn.de@localhost>]

On Mon, Sep 26, 2011 at 08:55:25AM +0200, Riccardo Mottola wrote:
> Hi,
> 
> On 09/23/11 12:19, Ignatios Souvatzis wrote:
> >
> >>Set ddb.onpanic?=1 in sysctl.conf. The kernel will fall into its
> >>debugger (ddb(4)) upon crash, print a backtrace with some useful
> >>info and then wait.
> >uh,,,, when it's remote, it's better to save kernel and corefile from
> >safecore, and analyze that (ddb.onpanic=0)
> 
> That's already configured that way in my sysctl.conf. Perhaps I have
> already the corefile? Where did savecore put it? Or do I need to
> configure savecore somehow first?

Should default on:

|marie is 105% grep savecore /etc/defaults/rc.conf 
|savecore=YES            savecore_flags="-z"
|                       savecore_dir="/var/crash"

assuming enough space in /var/crash (taking into account the number in
/var/crash/minfree) and enough space on the designated dump partition.

You'll need the netbsd.gdb from the same build as your then running kernel
to see the symbols; the kernel saved alongside is, normally, the running
one (from the time of the savecore) thus too symbol-lacking to be very
useful.

Regards,

        -is