Smarts NPM: Can I increase both the timeout and number of retries for BGP SNMP polling?

Products

• VMware

Issue/Introduction

Symptoms:


Can I increase both the timeout and number of retries for BGP SNMP polling groups used by Smarts NPM?
How do I change the BGP SNMP settings connect to the Smarts NPM  Domain Manager Admin Console

Environment

VMware Smart Assurance - SMARTS

Resolution

It is generally not recommended to increase both the timeout and the number of retries for a device. You generally should do one or the other (increase either timeout or retries) but not both. Increasing both will lead to very long waits before the final timeout. The default behavior is to double the timeout each retry, so the first requests waits 1.2 seconds, the first retry 2.4, the 2nd 4.8, the 3rd 9.6, the 4th 19.2, the fifth 38.4. If the device never responds that is a total timeout of 75.6 seconds.

If you need to change the Smarts NPM BGP settings for SNMP, do the following:

1. Connect to the  Domain Manager Admin Console
2. Connect to the BGP server and click on Polling and Thresholds.
3. Click on BGP Polling Groups on the tree and click on the BGP SNMP Settings

When to increase the timeout
Increasing the timeout is needed for long latency situations, such as a satellite link between one office and another creates a situation where you would need long latency. If a heavily loaded device does not respond quickly, then increasing the timeout might help.

When to increase the number of retries
Number of retries should be increased only if there is a large amount of packet loss, either because of the remote agent ignoring requests (which is legal in SNMP) or because of intermediate congestion causing routers to drop packets.