'How to optimally use the Windows machines as Jmeter Slaves for pushing higher loads

Greetings for the day!

I am currently facing challenges to induce higher load using Jmeter Distributed mechanism which I configured in AWS windows machines.

My machines are throwing Response code:Non HTTP response code: java.net.BindException Response message:Non HTTP response message: Address already in use: connect exceptions.

I am trying to calculate the foot print for the slave machines so that I can confidently communicate on how much load they can handle.

I am using a windows 2019 server, with a 16 GB RAM. While surfing through internet for answers I got the below link which says to change the registry settings to optimally use the ports. https://www.baselogic.com/2011/11/23/solved-java-net-bindexception-address-use-connect-issue-windows/

I am looking for suggestions from your end to execute more threads seamlessly (especially spikes using free from arrival thread groups) from the available resources I am currently having.

Please let me know if any further information is required from my end which can help me get better solutions.

Many Thanks for your support.

Regards, Vijay

jmeterperformance-testingdistributed-testing


Solution 1:[1]

  1. As per Adjusting TCP Settings for Heavy Load on Windows article you should:

    • create a DWORD value named TcpTimedWaitDelay with the value of 60 under HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\ Services\TCPIP\Parameters
    • create a DWORD value named MaxUserPort with the value of 32768 under the same key

  2. Make sure to follow JMeter Best Practices

  3. Make sure to monitor the networking and other performance counters of the Windows machine using either built-in Windows perfmon or via JMeter PerfMon Plugin

Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source
Solution 1 Dmitri T

Related Questions