How to detect which devices are causing poller overruns

Document ID : KB000023368
Last Modified Date : 14/02/2018
Show Technical Document Details

Introduction:    In the OneClick --Statistics Polling view, I am seeing poll cycles taking much more than 5 minutes.


Follow these instructions below to set the Poller advanced logging file to show poll times for individual devices:

1) Back up the $NH_HOME/sys/debugLog.cfg file
2) In the debugLog.cfg change:
   program nhiPoller[Net] {
           arguments       "-Dm poller -Df dtp -Dt"
   program nhiPoller[Net] {
           arguments       "-Dm poller -Df Ddtp -Dt"

3) First, ensure you remove any existing $NH_HOME/log/advanced/nhiPoller_Net.txt file, then in OneClick--Setup--Advanced Logging:

Turn on advanced logging for "Statistics Poller"
4) Let four polls pass to ensure we capture the errors.
5) Turn off advanced logging.
6) The $NH_HOME/log/advanced/nhiPoller_Net.txt will contain an line for each device such as:

<date> <time> [D,poller]   Device <uniqueDevId><port> completed <#> elems in <#> seconds with <#> backoffs, <#> skipped retries, and <#> skipped elems

<uniqueDevId> will be the uniqueDevId or "Hardware Id" of the device

7) To easily translate the uniqueDevId to a hostname or ipaddress:

In OneClick--Managed Resources--Elements:

Right click on Column headers--Select Fields: Add 'Hardware ID from the left panel to the panel on the right and click OK.

Search the 'Hardware ID' column for the uniqueDevID


If you need further assistance, open a case with Technical Support, and upload the resulting diagnostics file into the case from running:

nhGetSupportInfo -type "statsPoller"