No response from remote host on interface bandwidth monitor

[kcprospect]

We are consistently receiving 'No response from remote host' alerts when Nagios tries to monitor Interface Bandwidth on our Cisco Nexus 7010 devices.

This only occurs with interface bandwidth monitoring. All other monitors such as, CPU, Interface Operational Status, etc. do no produce this 'No response from remote host' alert.

I've ran continuous pings from my nagios server to the remote Cisco Nexus switch and have received no drops.

I've rebooted the Cisco switch to restart the SNMP process engine, but still no luck.

We have other Cisco and Juniper devices that are monitoring interface bandwidth without any issues.

Our service_check_timeout is set to 120.

I've attached a screenshot of the alert history we are experiencing.

If you have any suggestions or fixes or workarounds, it'd be greatly appreciated!

[tgriep]

Can you post the version of the plugin you are using and how the command and service check is defined so we can review it?
Most plugins have a timeout option, try increasing that to see if that resolves the issue.

[kcprospect]

We are using the check_bandwidth4 ... it's a version of the check_bandwidth with some minor modifications, but it uses the same process to call for interface bandwidth on our Cisco Nexus.

Our current timeout is set to 120s, what do you recommend it be increased to?

Thanks again for the quick reply!

[tgriep]

The service_check_timeout is a setting in the nagios,cfg time that stops a service check from running longer than 120 seconds, but I don't think that is the problem you are having.
The plugin that you are using probably has a much shorter timeout and that is causing the no responce message.
I need to get some details in the plugin as to where where you downloaded it from, it's version and how it is configured on your system.
Can you provide that info?
Also, if you run the plugin like the example below, it should printout the options for the timeout value.
Run this from the folder the plugin is installed in.

Code: Select all

./check_bandwidth --help

[kcprospect]

https://exchange.nagios.org/directory/P ... th/details

Really do appreciate your help ... as this is a mystery for us!

[tgriep]

Edit that script and change this line from

Code: Select all

ifindex=`/usr/bin/snmpwalk -v $version -c $community $hostname 1.3.6.1.2.1.31.1.1.1.1 | grep $interface | grep -o "\.[0-9]*\ "`

to

Code: Select all

ifindex=`/usr/bin/snmpwalk -v $version -c $community $hostname 1.3.6.1.2.1.31.1.1.1.1 -t 120| grep $interface | grep -o "\.[0-9]*\ "`

Than will increase the timeout to 120 seconds.

[kcprospect]

This is the code that our script is based on:

http://check-bandwidth.googlecode.com/s ... _bandwidth

[tgriep]

The second plugin has a timeout option you can use for your command.
-t, --timeout INTEGER
"Set the timeout value for communications with host via SNMP"

You can edit the command and add the following to it to give it a higher timeout.

Code: Select all

-t 120

This will increase the timeout for your check to 120 seconds from 15 seconds which is the default for that plugin.

[kcprospect]

Thanks for your feedback:

Would it be acceptable if I change the following 'timeout' value below under:

# command-line controlled options (set-up with default values)
%_options = (

'timeout' => 120,


=========== OR =========


If -t 120 is the recommended process, where exactly does this value/code get updated?

Thanks again!

[tgriep]

Doing either method would work for you.
If you edit the perl script, then you wouldn't have to edit command definition for the check_bandwidth.
If you use the -t option, you would edit the command definition in the commands.cfg and add the timeout value which would change it for all of the service checks that use that command.
Let us know the outcome of this.