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

HP-UX Handbook – Rev 13.00 Page 20 (of 38)
Chapter 08 Crash Dumps
October 29, 2013
that) they get decompressed automatically during the execution of crashinfo. This can take a while. Be sure to
have enough space left in the crash directory.
With the help of the webcourse mentioned above it should be possible to solve most of the
problems.
Anyway in some cases you might need information that is beyond the standard output of
crashinfo. In this case you can use one of crashinfo’s options or use the P4 debugging
environment to perform a deeper analysis.
At this point you have to decide whether to ship the debugging tools to the customer and provide
a remote connection for a HP RCE or to ship the dump to the Response Center, either via ftp
upload or on a DDS tape or CD-ROM by mail. Which of the above possibilities (remote login,
ftp upload, ship by mail) is appropriate depends on availability of remote login, the size of the
dump and the severity of the problem.
If remote login is not possible, ship the dump to the Response Center
If you choose to analyze the dump on the customers system, obtain the tools either from the
Ktools server (refer to the Additional information section below). Select p4, then shared, then
internet or via anonymous ftp from
ftp://tahoe.grc.hp.com/dumpreading/dump_analyse.tar.gz (HP internal)
Size is about 7MB.
Unpack the files e.g. below /usr/contrib/dumpreading on the system, where the dump is located.
Before starting you need to set some path variables:
export P4_ROOT=/usr/contrib/dumpreading
export PATH=”${P4_ROOT}/bin:${P4_ROOT}/p4:${PATH}”
export SHLIB_PATH=”${P4_ROOT}/bin:${P4_ROOT}/p4”
Either put the above lines in /etc/profile or simply source the included set_env file in order
to set these variables:
# cd /usr/contrib/dumpreading
# . ./set_env
If remote login is not possible, ship the dump to the Response Center
If you choose to ship the dump to the Response Center additional information from the
customers system depending on the type of the crash is needed.
Examine crashinfo output to determine which of the crash event types this is:
PANIC the system ran into an unhandable condition and paniced
TOC the system was hung and you TOCed it
SG TOC the TOC was initiated by MC/ServiceGuard via the safety timer mechanism
HPMC High Priority Machine Check. The crash was caused by a HW failure
Example:
# cat ci.out | grep "Note: Crash"
Note: Crash event 0 was a PANIC !