Down System Alert

[Coleman Blake_2]

I need a simple monitor to notify me if one of my systems has crashed. I've written scripts to do this before, but think there must be something already available since this is such a common need.

I don't need anything as elaborate as Big Brother/Sister or OpenView. I think EMS can be tweaked into doing what I need, but there may be something simpler that I've overlooked.

Coleman Blake
Peace Corps

[Steven E. Protter]

Though its a big tool and tough to learn, SCM might be a good choice for you. It monitors many levels of system functionality including which daemons are running. It also works for Linux servers.

Another choice is Openview TopTools. This is a snmp tool and you'll need to get that daemon runnning and configured with a public and community name.

Toptools runs on a PC and is quite picky about running on a HP Kayak PC. The tool is free, but can email page and do nice things like that.

You can do something very simple.

In a multi-server environment, you can run a simple script to check for the server via ping.

ping hostname -n 5 -n60
success=$?

if [ $success -ne 1 ]
then
take action to get hold of YOU
fi

SEP

[Dave La Mar]

Coleman -
As noted in the previous reply, a home grown ping script would work as well.
I think there was a typo in the suggestion though.
I beleive it should be:
ping hostname -n 5
success=$?

if [ $success -ne 0 ]
then
take action to get hold of YOU
fi

Best regards,

dl

[Bhuvaneswari Selvaraj · Accepted solution]

Hi,

You can use OnlineDiagnostics which is actually uses EMS.

You need to install OnlineDiagnostics on your home system and also the system you want to monitor. OnlineDiagnostics has a monitor /usr/sbin/stm/uut/bin/tools/monitor/sysstat_em which actually monitors the system status. After installing the bundle on both the systems, using monconfig (this is a simple tool, and is almost similar to EMS in taking parameters) on your home system, mention the system that you want to monitor and also specify how you want to be notified. When the monitored system goes down, you will be notified. Hope this helps.

[Coleman Blake_2]

Thanks for all the replies and sorry for the late response. I was hoping to analyze my results and provide some more information, but have been OBE'd.

I didn't want to use the ping based solutions because a momentary network glitch could get me paged at 0-dark:30.

I downloaded SCM and hope to get a chance to learn about it. It's a big, powerful package and will take some time to learn how to use.

I used an OnLine Diagnostics/EMS approach.

The configuration file for sysstat_em is /var/stm/config/tools/monitors/systat_em.cfg. At the bottom of this file is a commented example of systems to monitor. If you add systems there, you can configure alerts through SAM.

I shutdown all the systems before Huricane Isabelle and got lots of pages. They are the results I haven't been able to analyze.

I haven't been able to get send_test_event to work yet, but that may be my problem.

[Jeff Schussele]

Hi,

send_test_event can only send an informational level event. So if you don't have the info level set to page out or email, then you won't receive the notice. It will, however, log the event to the event.log.

Rgds,
Jeff