Submit a ticketCall us

Get a crash course on Network Monitoring delivered right to your inbox
This free 7-day email course provides a primer to the philosophy, theory, and fundamental concepts involved in IT monitoring. Lessons will explain not only how to perform various monitoring tasks, but why and when you should use them. Sign up now.

Home > Success Center > Network Performance Monitor (NPM) > No More Net-SNMP Nodes (Machine Type List)

No More Net-SNMP Nodes (Machine Type List)

Table of contents
Created by Ian Rossiter, last modified by MindTouch on Jun 23, 2016

Views: 16 Votes: 0 Revisions: 4

Overview

This article describes how to change all non-Windows servers from being catagorized and put in the folder of "net-snmp" in the machine-type list. With the new NPM 11.5 feature of being able to assign standard elements to different OIDS, this is a thing of the past.

Environment

NPM 11.5x

Steps

 

Objectives

  • Set up a script on the target node (this will exist on all your target nodes) that responds to the command "osname" with the actual distribution and version.
  • Edit the net-snmp config to run that script when a specific OID is polled
  • Assign that OID in NPM to replace the existing machine type.

Follow the following sequence to achieve these objectives:

 

Get Scripting

  1. Depending on your distribution, you might use uname, or lsb_release, or some other method. The upshot is that you want the command to output the OS name. For example,
    • Ubuntu: uname -v | awk '{ print $1 }'
    • SuSe: lsb_release -d | awk '{ print $2, $4, $5, $6}'
  2. Create a script and roll it out to all systems. Once you have a script that works (even if it is different for different Unix/Linux flavors), create a script (or scripts) that do that command. For example,
    #!/bin/bash
    uname -v | awk '{ print $1 }'

    then place this script in the same directory on all systems.
    It does not matter if the SuSe version does different commands than its Ubuntu counterpart. It matters that it is always called /usr/bin/osname.sh.

 

Update your SNMP configuration

  1. In the /etc/snmp/snmpd.conf file, add the following line:
    extend osname /usr/bin/osname.sh
    restart the snmp service each time you make this update.
  2. Perform multiple tests.
    This is going to involve running "snmpwalk" a lot to make sure you can get SNMP to display this new value. Start with:
    snmpwalk -v2c -c <rocommunity> <hostname>
    and once you do not get errors, try:
    snmpwalk -v2c -c <rocommunity> <hostname> NET-SNMP-EXTEND-MIB::nsExtendObjects
    and finally:
    snmpwalk -v2c -c <rocommunity> <hostname> NET-SNMP-EXTEND-MIB::nsExtendOutLine.\"osname\".1
  3. Test some more.
    Once that works, you need to get the OID of your new object. While you can use "snmptranslate" for that, you may be better off using the snmpwalk.exe utility built into SolarWinds. This also verifies that you can walk this new object from the SolarWinds poller. It is typically found in C:\program files(x86)\solarwinds\orion\snmpwalk.exe.
  4. Roll the change out to all systems.
    Now that you have the snmpd.conf file working, roll it out to all your other *nix systems as well.

 

Add the poller to NPM

  1. Go to Settings > Manager Pollers, and create a new poller.
    1. Technology is Node Details.
    2. On the next screen, select MachineType and define a new data source.
      1. Cancel out of the Set a constant value.
      2. Click Select an OID.
      3. Scroll to the bottom and select the Looking for an OID... box.
      4. Give the OID a name (like "osname").
      5. Enter the actual OID you discovered in step 4. It may take some trial and error since the snmpwalk is going to provide a very long number. You will take a portion of it.
    3. Pick a new SysObjectID.
  2. Apply the new poller to the appropriate *nix devices in your environment.
Last modified
22:52, 22 Jun 2016

Tags

Classifications

Public