marsh@umt.UUCP (Paul Marsh) (05/22/87)
I'm not an 'adb' expert {yet}, but days like today make me wish that
I was. We experienced a number of system crashes which took most of
the day to track down. Has anyone compiled any information that would
help in looking through a crash dump? A few years ago I attended a
TOPS-20 session at DECUS, where the speaker(s) had compiled a "cook-
book" for looking at TOPS-20 dumps, which was very helpful. Something
like that would be nice, if it exists. Is there any information that
details the possible causes of the various "panics:"? Thanks in advance.
UUCP : ... ! ucdavis ! umt ! marsh
US Mail : Paul Marsh, University of Montana, Computer Center
Missoula, MT 59812
phone : (406) 243-5455pdb@sei.cmu.edu (Patrick Barron) (05/24/87)
In article <680@umt.UUCP> marsh@umt.UUCP (Paul Marsh) writes: > [...] Has anyone compiled any information that would >help in looking through a crash dump? > [...] Is there any information that >details the possible causes of the various "panics:"? Thanks in advance. Assuming you've loaded /usr/doc on to your system from the distribution and uncompressed it, look in /usr/doc/kdebug, which contains a paper called "Using ADB to Debug the Unix Kernel" (or words to that effect). It has a few helpful hints (probably the most useful is that doing *(scb-4)$c while ADBing a crash dump will give you a stack traceback of what was running at the time of the crash). I don't think this paper is in the hardcopy document- ation set for Ultrix, but I could be wrong. A very few of the possible panics are detailed in the man page for crash(8). However, that's just the tip of the iceberg - there are really more possible panics that you'd want to think about. If you've got a particular panic you're curious about, the best thing to do is look around in the sources (assuming you've got them) and find out where that panic is generated. --Pat.