One of my poller show error about poller over run. Even we Increase PollQueueLimit again, this time to 15000. Kindly, share any information. As it seems to be known issue.
I'm guessing you are referring to an update to this query to show the SNMP Loss polling failures instead of overruns.
mysql nms2 -t -e "select from_unixtime(event_tstamp), event_descr from event_log where event_descr like '%overrun%' order by event_tstamp desc " > D:\overruns.txt
Just change the event_descr
mysql nms2 -t -e "select from_unixtime(event_tstamp), event_descr from event_log where event_descr like '%100% SNMP Loss%' order by event_tstamp desc " > D:\Loss.txt
Can you clarify which product you are having this issue on please.
Including the product in the title or adding the relevant category will ensure that the question is routed to the correct teams for review.
The product is CA-Netvoyant version 7.1.108
Hi NetVoyant community, can someone assist with this question please?
Yes polling overruns are known to happen for various reasons.
An overrun is when NV failed to complete polling for all the poll instances for the dataset within the poll cycle. It only takes one missed poll instance to generate the message and the message give no indication as to how many poll instances were missed. It could be there was not time to even try polling the poll instance, or it could be the SNMP request was sent but the device did not respond.
The most common are:
an overloaded poller
too many poll instances being polled in too small of a polling cycle, common with a one minute poll cycle
devices which are poor responders to SNMP polling
Increasing PollQueueLimit will let more polling get done at the same time, however it will not help if the system has insufficient CPU or memory, or there are problems with specific devices.
A single problem device may have a significant impact on overruns. A single device can have hundreds or even thousands of poll instances and for each there is the time out and retry settings which extend the amount of time the device will waste.
Setting the Polls service to Logging Level: Normal (Verbose) will result in details messages in the event log showing which devices are getting the polling overruns and how many are pending.
These tech tips offer recommendations for devices with polling failures.
Tech Tip - Polling Failure - 100% SNMP Loss
Tech Tip NV - 'No Data Returned' Polling Failures
Tech Tip: NV - Event Log Filtering
Thanks for sharing information. it really give us idea about the issue. I am still tryinh to fix the filter result. Kindly, share some query to get all Polling Failure - 100% SNMP Loss.
It works I found a big list of devices instance failed. Once again thanks for your support and time.
Retrieving data ...