Debugging a hanging 5.4 system

Michael R. Wayne wayne at
Wed Jan 25 10:40:11 PST 2006

We've got a system running 5.4-RELEASE-p8 that occasionally
hangs while dump is running.  I suspect it's something to do
with the SCSI controller hanging:
   ahc0: <Adaptec aic7896/97 Ultra2 SCSI adapter>
because the machine remains pingable and I can hit <CR> on the
serial console and keep getting login prompts until I enter a
username at which point the machine ceases to respond.

I've compiled a kernel with 
include		GENERIC
options		KDB
options		DDB

in preparation for the last crash.  There is a portmaster hooked
to the serial port so that I can telnet to it and break to the
debugger.  Obviously, I will not be able to create a crash dump.

My question is what specific DDB commands I should use when the
system is hung so that I can generate sufficient information to
create a useful PR.  Any other hints/tips would also be appreciated.

/\/\ \/\/

More information about the freebsd-hackers mailing list