Nagios Support Forum

We have been monitoring several Cisco switches as hosts just to see if they are reachable. Now we wanted to monitor the bandwidth of a few ports.

I trye'd to add the port bandwidth graphing via the monitoring wizard, but it pulled all the ports ( this is a switch stack with 8 52-port switches ) and it became very difficult to find the port I needed ( the ports are identified by port # rather then name ), also all the ports a checked by default and to uncheck all the ports ( 416+ ports) is time consuming to say the least.

So I am trying to configure a single port manually.

I configured mrtg to monitor the port I needed, then created a service using the check_xi_service_mrtgtraf command and added the xiwizard_switch_port_bandwidth_service template.

It is pulling the data as expected and displays it in the service status table, but it's not generating any graphs.

Am I missing a configuration step?
Is this covered in a manual ?

I need to monitor a few other ports on different switch stacks, so I would like to do it manually on a per port basis rather then go through the wizard.

Thanks
Sal

Make sure you copy the settings generated by the Wizard carefully. The Wizards should not be able to effect changes at a lower level then the Configuration GUI.

As for port Name VS Index make sure you have the correct MIB(s) loaded for the oid(s) you plan to use. It would not work to simply change the port Index into a string as a Name, you will have to find the correct oid(s) you plan to use.

snmptranslate can help here.

Thanks for the reply.

Not sure which config files you are referring to, I looked at the config files created by the wizard and only saw the mrtg config are there others?.

As far as the port ID's go I thought Nagios was using the MRTG configmaker utility which automatically pulls the port names by specifying ifdesc=names in the argument line.

This is becoming a little frustrating, I have used Nagios for many years, we moved to NagiosXI because of the integration of all the different tools ( Cacti, MRTG, etc..) but I am finding it difficult to understand how to customize the system.

Sal

I believe the issue I am having is that there are no files being created in the perfdata directory of the device.
In the perfdata directory there is one pair of files _HOST.rrd and _HOST.xml .

How do I get the perfdata for the port bandwidth service ?

Thanks
Sal

Are you sure you're using the same command definition that the wizards use for monitoring bandwidth? It sounds like a different command is being used, as the proper command will run the right plugin that generated performance data (for later graphing by PNP).

FYI, we've modified the switch monitoring wizard to allow you to change the port name. This will be in the upcoming 2009R1.1G release in the next day or so.

I look forward for the next release the port names will be a big help.

Will the next release also include a method to either check or clear all the interfaces found by the wizard?
That would be a great feature for switches as they can have hundreds of ports and only a few need to be monitored.

Thanks for the reply.

Sal

I resolved the issue after I double checked the config as you suggested.

I rechecked the config and there were two difference's:
1> the wizard was using the M switch to display in MegaBits and I was using B to display in bits
2> The wizard was using 50,50 and 70,70 to specify warning and critical, I was using a single number 50000 and 70000


Thanks for the help
Sal

Glad to hear you got this figured out!