From: Cian O'Sullivan (Cian@logic.bm)
Date: Mon Apr 29 2002 - 08:37:11 EDT
Thanks to those Pat O'Brien, Ashraf Bakar, and once again to Selden E. Ball.
Basically there are three steps you can take, of which each has an increase prerequisite of knowledge to understand.
1. Look in /var/adm/messages. I noted that if the crash means that the system could not write to the /var/adm, it will write the messages in the /var/adm on reboot.
2. Run the uerf command. (uerf -R -o full | more ) will give you a long and comprehensive list of the /var/adm/binary.errlog It is exactly what I needed.
3. Look in the /var/adm/crash for the core dump files. the crash-data file should give you the actaul stack trace that caused the dump, however it did not have this info for me.
Thanks again.
Cian O'Sullivan
====ORIG QUESTION =====
Subject: RE: How do I debug why a box crashed, and both kernel file
corrupted.
I like to start before the reboot. did the system crash and dump. I
usually review different registers, and make a determination to manual crash
based on results. during reboot does the system copy the dump out
automatically. what is in the crash data file. and then there is the logs.
logs is a complete subject, but I keep them local and remote. MANy time the
remote logs are more benifical than local ones. when a system crooks in the
night time and is down, i review logs on the remote server before arriving.
Other time the renmote logs are the only logs in existance. but these steps
take planning before a event to capture.
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:39 EDT