Submit a ticketCall us

Announcing NCM 7.7
With NCM 7.7, you can examine the rules that make up an access control list for a Cisco ASA device. Then you can apply filters to display only rules that meet the specified criteria, order the rules by line number or by the hit count, and much more.
See new features and improvements.

Home > Success Center > Network Performance Monitor (NPM) > Topology Polling causing high CPU load on network nodes

Topology Polling causing high CPU load on network nodes

Overview

Customer experiences high CPU on their network nodes after adding them to Orion for monitoring. 

Environment

All NPM versions

Cause 

This issue is caused by topology polling. In this process, the node must gather the mac address table from all interfaces and compile it before sending an SNMP response back to the NPM server. This can take up a lot of resources and put a large load on the CPU. It is suggested that you turn off topology polling for the affected nodes, and wait for about 15 to 30 minutes and see if the CPU load decreases. 

This is a common issue in Cisco routers as well as Core routers that perform BGP which have larger routing tables and large ARO tables in Enterprise LANs. 

Note: SNMP is a low priority on a router's CPU stack which means, you will drop the SNMP packets. For more information on this, you can visit Cisco CMTS Universal Broadband Router Series MIB Specifications Guide 12.2SC - Chapter 2, Configuring SNMP and MIB Support  [Cisco uBR10000 Series Universal Broadband Routers] - Cisco Systems.

Resolution

Disable topology calculations:
1. Log in as an administrator to the Orion Web Console on the SolarWinds server. This one is responsible in polling the object for which you want to disable topology calculations.
2. Click Settings in the top right corner of the web console.
3. Click Manage Nodes in the Node & Group Management grouping.
4. Check the node for which you want to disable topology calculations, and then click List Resources.
5. Clear all Topology checkboxes.
6. Click Submit.

 

To edit the polling interval:

1. Click Settings > Polling Settings and look at your Default Topology Poll Interval. You can increase this to 60 or 120 minutes (or more) to help alleviate the load on the switch. This is only optional if you choose to continue using topology polling. 

 

Note: If turning off topology polling does not reduce the load on your node, edit the affected node in the web console and change the polling from SNMPv2 to SNMPv1. The SNMP v1 protocol does not include the GetBulk Request that is in SNMP v2. When NPM sends a GetBulk request to a device it has to build a large dataset before it sends a response, which can cause high CPU use. You might have to restart the SNMP service on the switch after changing from v2 to v1 to clear out the SNMP input queue. Once this is done, CPU use should go back to normal.

 

 

See this thwack post for more details..
http://thwack.solarwinds.com/thread/51415

 

 

Last modified
15:10, 10 Mar 2016

Tags

Classifications

Public