We ran into an issue with an SNMP device that has been switch to SNMP Shaping/Rate shaping, due to this change the device is now polling significantly slower up 20x times slower. Unfortunately, this policy will be applied across all devices at the customer, Is there any way we can combat the performance issue due to SNMP Shaping mode?. The driver currently uses a blend of native SNMP polling and Qaction SNMP polling.
the picture shows the configuration of the device
Note we have also collected Wireshark capture for a device with the mode enabled and one with the mode disabled if this is needed to analyze further.
From the configuration, it seems that if SNMP requests come in at a rate higher than 512 packets per second, they will be dropped.
This will have a significant impact on your polling performance:
1. You can only poll at a very low data rate
2. Exceeding the limit, will cause retries which will slow down the communication even further.
Using the 'show policy-map POLICY-COPP interface yourinterface' command on your switch/router you should be able to see how many packets were dropped.
Using the Timer Base on your element, you can then slow down the polling until no drops occur anymore.
To have critical information coming in, in a timely manner, you will need to rely on e.g. SNMP traps.
Hi Aldrick,
The best way is to perform the check on the device itself.
If this is not possible, then you’ll have to fall-back to Wireshark.
In Wireshark, you can see the this info through the statistics -> IO Graph menu.
Thank you Ive, Do you know if there is a way we can check the number of packets sent between DataMiner and the device?