Server crashes when not in use.

UNIX Solaris
1

Production server crashes when there is no traffic on it. Can only recover by going to ALOM and reboot the server. This seems to happen about the same time every month. The only good thing is the server is not in production at the time of the crash.
I have been unable to locate any information in the logs. Cause could be from remote backups or security scans but can't prove it. All hardware testing shows no problems.
Nothing shows in var/adm/messages files.
I would like to know how to setup logs that may save information. Any information will help.
System is a Sun V210.

Thanks,

2

What version of Solaris? Does anything show up in prtdiag? I would check for a cron job since it goes down about the same time each month. Is pmconfig running? If so check /etc/power.conf, but I can't imagine this causing a problem only once a month. After you bring it up, run "last" to see who was logged on when it went down. Also check /var/adm/sulog to see if anyone was using su when it went down.

3

Running Solaris 10.
This is what I checked using your suggestions.
pmconfig = does not show up in ps -ef
sulog = no info the day and time of last crash
last = Mon Sep 18 (most current date).

Can prtdiag be run while the server is in production.

Thanks

4

Are you getting a crash dump from the kernel?

Or is the server not actually crashing but just hanging?

5

Nothing on the screen.
Not getting any crash dumps.
The system looks like it is hung.
Have to go to alom sc> to reset it.

I was thinking of maybe opening a telnet session to the box and leaving a alarm window open. Maybe it will catch something when it does it again.

Thanks

6

prtdiag can be run while the server is in production and it's a good idea to do so every now and then. It's easy on Solaris 10. As root, just run prtdiag -v

--edit--
I just plugged a serial cable into the management port of a V210 to look around the Service Processor (aka ALOM). I see a consolehistory command that might be worth trying the next time it happens before you reboot. And yes I am in ALOM while the box is up and running.