Thanks everybody who helped, this is what I have been told.
A kill -SEGV or 'gcore' will cause a running process to core dump.
To cause the kernel to panic, use adb. Find the symbol for the / and
basically over write it with 00000's, next time the machine needs to use the
filesystem (even to access a dev entry) the machine will panic.
savecore needs to be enabled on 2.6 for the machine to save the dump when it
comes back up again, /etc/rc2.d/S20sysetup.
/usr/bin/savecore -d <path to directory to put the dump>
The -d is undocumented, I believe. This works best just
after a reboot, before swap has gotten stirred up much.
Out of interest, we always have a spare partition capable of holding a
system crash dump in the event that one doesn't get written.. ;-)
Thanks go out to
Mark Sowerby
Darren Dunham
Don
*** ORIGINAL QUESTION ***
>
> My real question is, how can I cause a crash/core dump from the command
line
> of either the whole box or an individual process.
Thanks again.
Andy Holmes
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:33 CDT