Subject: savecore not,uh, saving core
To: None <netbsd-users@netbsd.org>
From: Rasputin <rasputin@idoru.mine.nu>
List: netbsd-users
Date: 11/27/2002 10:47:21
My sparc is panicing horribly with the latest kernel.

MCSE-level diagnosis (
"try connecting again?" <BANG> "hmm, that shouldn't happen"
) says it's allergic to wireless Mac clients, but that means I
can only apply MCSE-level fixes, and the reset switch is
taking a hammering.

I've caught it crashing to a serial console, but after thousands of lines
it ends with 'watchdog timeout' 'stack underflow' etc.

savecore=YES

Is in rc.conf, and I can see 'checking for core dumps' when it boots.
/var/crash has 700Mb free (minfree is 1024), total RAM is 320Mb w/500Mb
swap, so I should be getting something?  Is that enough swap?


Is it the serial console causing this behaviour? If so, how can I get a
coredump?

And if anyone has useful debugging techniques for wi(4) on sparc/nell(4),
now would be a good time to mention them :)

Thanks folks.

-- 
Rasputin :: Jack of All Trades - Master of Nuns