>        We have had several kernel panics and would like to capture some
>information as to the location of the crash point.  How do we set our
>machine up to panic and then dump on boot so the crash can be analyzed.

It would be better to include the kernel debugger than to mess around
with savecore.

add

options KDB

to your config file.  Then panics will give a kdb> prompt.  Just don't
run X11 on the console.

But if you must have an image, leaving out 'options KDB' should allow
the system to write a dump to your primary swap space (which must be
big enough to hold it if you are running unpatched 1.0).  You will see
a series of integers and the hard drive will run for a while during
the dump.  The savecore program will copy the dump to /var/???/crash
when the system boots.  Use "savecore -c" to mark the swap as not
having a dump anymore.  Run "crash" on the image in /var...