HP-UX HB v13.00 Ch-08 - Crash Dumps

HP-UX Handbook – Rev 13.00 Page 19 (of 38)
Chapter 08 Crash Dumps
October 29, 2013
Analysis of the dump
A complete analysis of a crashdump requires deep internal knowledge and much experience.
That would certainly go beyond this document. Here I'd like to explain how to use the utility
crashinfo in order to narrow down the cause of the crash.
If you like to examine the dump by yourself, please refer to the excellent online webcourse
offered by the Expert Center. This course should be considered as starting point for any dump
analysis. Whenever you deal with a crashdump i recommend you to visit this site. In most cases
you should be able to find a solution. Links to all available dump reading tools are included.
http://teams3.sharepoint.hp.com/teams/esssupport/InsideESSSupport/InsideWTEC/HPUX-
KERNEL/crash/FirstPassWeb/index.htm (HP internal)
About the crashinfo utility
crashinfo is an executable that is based on libp4, the library of the P4 kernel debugger. It replaces
the old whathappened perl script that was based on the Q4 kernel debugger. P4 and crashinfo are
much more powerful and advanced than Q4/whathappened. The P4 debugger is based on the
korn shell (ksh88) which makes it comfortable to use and the libp4 library.
p4 and crashinfo can be performed on a dump (by executing it from within a crash directory) as
well as on a live system (by executing it not within a crash directoy). The latter can be useful to
examine kernel structures when the system is e.g. not completely hung.
crashinfo is smart, depending on the type of the crash (PANIC, TOC, SG TOC or HPMC) it
prints out the appropriate structures. It also reacts to certain conditions e.g. system low on free
memory, spinlock panics, etc. and prints out the necessary data.
How to obtain and install and execute crashinfo
Use the standalone version to perform a quick check
Obtain the static (standalone) version of the crashinfo binary from the Ktools server (refer to the
Additional information section below). From there the tool can be sent to the customer by email
or pushed to an external ftp server. Size is about 800K.
Store the static crashinfo binary e.g. at /usr/contrib/bin/ on the affected system.
This version should be used to perform a first quick check of the dump.
To get a fingerprint of the dump simply run the standalone crashinfo without options from within
the crash directory:
# cd /var/adm/crash/crash.0
# /usr/contrib/bin/crashinfo >ci.out
NOTE: Should the chunkfiles (image.n.m for a UX 11.X crashdump) be compressed (the suffix .gz indicates