log files to be monitored

[Vaish Raj]

Hi All,

I am new to unix/HP administration.

1)I want to know the log files i have to monitor hourly/daily/weekly to check health of system and prevent any downtime.

Can someone also explain best methods/options to maintain these log files and what errors to look for in each log file.

Detailed explanation is appreciated.

Thanks

raj

[Sridhar Bhaskarla]

Hi Raj,

You can do with scripting but you have to spend quite some time and effort. Look at HP's porting site to see if any of the tools may be useful.

http://hpux.cs.utah.edu/hppd/hpux/alpha.html

My best advice to you is to enable EMS. EMS can monitor hardware and some software components and send alerts to mails (or any other means) accordingly. Here is the information on it.

http://docs.hp.com/hpux/onlinedocs/B7609-90022/B7609-90022.html

Tools like Openview VPO etc., can help monitor the system log files like syslog.log etc., but you will need to purchase them.

-Sri

[Bharat Katkar]

Hi Raj,

1.As far as OS logs are concerned they are in
/var/adm/syslog/syslog.log
2. Another one is shutdown log which talk about why the system shutdown happen or who was responsible for shutdown. That is in /etc/shutdownlog
3. Also if you have any application running it may it own log files which again you need to monitor. Mostly all log files are under /var directory.
4.Other than log monitoring you can even monitor filesystems by "bdf" command which tell you about filesystem Usage.
5. You can also view GSP log for any hardware warning or faults. That is indicated by the Alert Led which is normally on the front panel of your server. Use B combination to go to GSP and type SL to view system logs.
6. You can separate out different type of log messages by configuring syslogd daemon. see man syslogd.

Keep reading more about HPUX and understand the way it works.

Hope this helps.
Regards,

[Geoff Wild]

Main log is /var/adm/syslog/syslog.log

Use OVO to monitor - or if you can't afford it, try BigBrother:

http://www.bb4.org/

Best methods? hmnnn...have a look at http://docs.hp.com

Rgds...Geoff

[Bill Hassell]

There are a few that always need monitoring, mostly in /var/adm:

wtmp btmp syslog/syslog/log

But then the list gets highly dynamic depending on what you do with your server. So analyze /var/adm on a weekly basis:

du -kx /var/adm | sort -rn | more

Or you can take the easy route: run SAM and it will show you the logfiles and allow you to trim or zero them.