Tech Tip - Polling Failure - 100% SNMP Loss

Document created by DavidM Employee on Jul 16, 2014Last modified by DavidM Employee on May 15, 2015
Version 4Show Document
  • View in full screen mode

In earlier builds of NV the message was 'No response before SNMP timeout' but was changed to '100% SNMP Loss'.

In NV this is controlled by the SNMP Timeout and SNMP Retries settings for the device, with the defaults at Timeout 5 and Retries 1.

Typically the number of polling failures due to 100% SNMP Loss can be minimized by adjusting these values but there are several things to consider.

What the polling rates are.
If the device is known to have high SNMP latency.
If the device has hundreds of poll instances.

If the SNMP Timeout and Retries settings are too high then NV will still be waiting by the time the next polling cycle starts resulting in overruns typically when 1min. polling is used making a very small polling cycle window, or there are hundreds of ifStats poll instances (timeout + retries) * # poll instances can also easily overrun a 5min. polling cycle.

For devices with known latency higher values will be needed.

Additionally it may help to change the length of the polling cycle for specific devices.
To do this, select the data set such as Interfaces > right click Select Polling Rate, in the drop down list will be the polling rates for the data set.

As a general guideline typically setting Timeout to 3 and Reties to 10, or 4 and 15 works well.

 

Note: The more retries that are done the more the 100% SNMP Loss stats will increase.

 

Verify the device is not using SNMPv1.

 

If the above does not resolve it try adding the isGet property to the device.

If it does not help remove it.

At the Poller Console

Groups tab

Select the device

Properties

Add

Name: isGet

Value: 1

click OK

Rediscover the Device

 

This document was generated from the following discussion: Polling Failure - 100% SNMP Loss

1 person found this helpful

Attachments

    Outcomes