Monitor 'link down' events

[Weaver]

Hi,

I'm testing HostMonitor in our company as a backup/second monitoring solution to HP Openview Operations. (I like it actualy more then openview )

I would like to monitor simple link up/down events for snmp enabled network hardware such as routers to our branch offices & uplink ports from the switches to the backbone.

When I make a test "Traffic Monitor" and let HM generate a list of the ports there is a column 'status'. I just want to have the test set to 'bad' if the status goes 'down' for a port. Can't be so hard but I can't find it neither in the manual or the forum. Anybody?

[edit]
To be more specific, I do know that it is possible to receive an SNMP Trap for link down events but then I will need to configure all devices with the ip of the hostmonitor if I'm right. I need to know how hostmonitor can poll the device itself. Otherwis how can I receive it anyway if the connection of a device to the backbone is down.
[/edit]

[KS-Soft]

When I make a test "Traffic Monitor" and let HM generate a list of the ports there is a column 'status'. I just want to have the test set to 'bad' if the status goes 'down' for a port. Can't be so hard but I can't find it neither in the manual or the forum.

Traffic Monitor test does not check status of the port (yet). However its easy to implement such test using SNMP Get test method.
Use OID 1.3.6.1.2.1.2.2.1.8.<index> to check status of the interface (instead of <index> you should use index of the interface).
If you want to start alert when interface goes down, use following condition "Alert when OID [1.3.6.1.2.1.2.2.1.8.<index>][is <> from][1]

Regards
Alex

[Weaver]

Thx! I'm not really familiar with SNMP but working on it I will try it out later this day.

[Weaver]

It works fine, except for one detail, if a link goes down the status changes to 'unknown' instead of 'link down'. So I guess when an other exception occurs (one powersupply blows up for example) it will also report 'status unknown'. Am I right? If so, is there a workaround? I'd lik every part of the network hardware monitored by a seperate test.

[KS-Soft]

It works fine, except for one detail, if a link goes down the status changes to 'unknown' instead of 'link down'.

I assume system where HostMonitor is installed and device that you are monitoring connected by interface that you are checking? In this case there are no sense to use special SNMP test, use simple ping test.

Regards
Alex

[templetonmd]

We have been using Hostmonitor for three years now and recently began using HP Openview. I am curious if you (or anyone) has any experience with having Hostmonitor alerts hit the HP Openview console and/or reporting Hostmonitor data (sent to a SQL Server DB) through HP Openview.

[KS-Soft]

I see only one way to do this - HostMonitor may send SNMP messages to HP OpenView using "SNMP Trap" action.

Regards
Alex

[timn]

Thx! I'm not really familiar with SNMP but working on it I will try it out later this day.

Over the past couple of years we've become much more dependent upon monitoring via SNMP.

We use HM to monitor traffic, status of fans, power supplies, temperature probes, as well as input/output voltages, output amperage, load %, etc on our UPS units.

Currently working on how to monitor our backup Caterpillar generator.

It all boils down to knowing the number of the OID you need to request -- and the range of values it can take on. Eventually, we found we needed a companion tool for exploring the various MIBs on our network devices.

There are many such tools available - for both Windows and favors of Unix. Some are free; other are low-cost shareware; others are retail products. We tried a few of the free Windows-based tools but eventually settled on OID View from Bytesphere. Cost is about $280 for single user license. For us, it makes a great companion to HM.

I would urge you to explore a few of the free tools first to see if they would be adequet to get you going.