My Oracle Support Banner

How to Optimize CPU Performance to Fix Napatech Buffer Drops (Doc ID 2018751.1)

Last updated on OCTOBER 09, 2020

Applies to:

Oracle Communications Session Monitor - Version 3.3.90 and later
Information in this document applies to any platform.

Purpose

If you experience packet drops in the Napatech host buffer and have already increased the buffer size to 1024 (see <Document 1934414.1> on more details), then this article can be used to optimize CPU usage in the system. This enhances the performance on the "rat" subsystem that reads data from the Napatech buffer, and if the data is read fast enough, then overflows/drops in the host buffer can be avoided.

Note: this document applies to version 3.3.90.1.0 or higher

Note that this document should only be used if recommended by ORACLE Professional services/TSC Support

 

Troubleshooting Steps

To view full details, sign in with your My Oracle Support account.

Don't have a My Oracle Support account? Click to get started!


In this Document
Purpose
Troubleshooting Steps
 1. Disable munin
 2. Get to know the CPU setup
 3. Get to know the Napatech NUMA node usage
 4. Check current CPU IRQ load
 5. Assign the CPUs
 6. Wrap up and restart  
 7. Check that the CPUs are used correctly
 8. Next steps if performance is still not satisfactory
 Replace sniffer type "napatech_direct" with "napatech"
 Disable irqbalance and get full control of IRQs
 See also other performance improvements
References

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.