AIX

Hello All,

Critical AIX production box crashed/rebooted while our team is working on it and we need to generate a detailed report for that, below are few questions that need to be included in the report. (We are System Administration team and everyone in our team has root access via sudo as well as root password, that is our main application server.)

1. Who is doing what? at the time of crash/reboot?
2. Was it user initiated roboot?
3. If user initiated then how can we substantiate?
4. What are the commands that need to be run to get our desired out put and why are we running those commands?

Regards,
Lovesaikrishna

Check system logs and user shell histories for a start.

last command will show you if it was a crash and who was logged in at the point in time
errpt will show you if the server had any problems
if its an oracle box - particularly rac - check the oracle logs if the node has been brought down by oracle. If you run nmon in data collection mode, check if you exceeded maxpin or if you were running out of paging space.

If you have a crash dump send it to IBM - they can determine what was exactly causing the crash

regards
zxmaus

Hi

The best way to find the root cause analysis is to log a PMR with IBM. Recently I met a SAP application server crash which is running in AIX. IBM may ask you to execute a few commands like

Code:

# snap -r
# snap -gbc

Once you send all the output they need they will come back to you with their analysis.