Jmeter- distributed testing, save separate results for each slave - jmeter

I have 1 master and 2 slave machines setup. I want to create two separate summary report, one for each slave.
I know the question in duplicate of this question, but the solution provide does not work for me.
JMeter Slave Reporting in Distributed Testing
In my test plan I have summary report and have added "filename"(or complete path) under write results to file.
When I execute my test plan, result file is saved in master system with combined results of both the slaves. How can I get the response times of the requests executed with each slaves separately?

If you add a Listener to your Test Plan and configure it to save results into a file it will create a results file on each slave machine so once your test is done you will be able to fetch these results file from slaves and perform required analysis.
Assuming your configuration is good you should have 1 cumulative result on master and 2 separate results on each slaves.
Another option is using __machineName() or __machineIP() function as a sampler label prefix (or postfix) so you will be able to distinguish samplers coming from different slaves in the main .jtl results file. This approach will allow you to avoid unnecessary disk IO overhead during test execution required to generate an extra file with the results.
Once you test finishes you will be able to split main .jtl results file using Filter Results Tool to generate separate results files per slave basing on slave machine hostname or IP address.
You can install Filter Results Tool using JMeter Plugins Manager:

Related

Jmeter- no results obtained using jmeter distributed system

Im running Jmeter test with one master machine and slave system. The issue is Im not getting the results from the slave system, even though I could see the test has started and finished in slave system.
I tried running another Jmeter test with a different script & same configuration and fortunately Im able to get the results from the slave system.
Im not able to understand why the first scenario is not working and where exactly is the issue. Kindly help me out.
From your screenshots I can only state that the test didn't produce a single SampleResult because it didn't execute any Sampler, the most common reasons are:
Your test is using CSV Data Set Config and you forgot to copy the CSV file on slave machines. JMeter master only transfers the .jmx test plan to slaves, everything else needs to be copied manually
Basically a subset of point 1: your test plan relies on a JMeter Plugin which is not present on the slaves. If this is the case - install the missing plugin(s) using the JMeter Plugins Manager
The exact failure reason can be found in jmeter.log file on master machine and jmeter-server.log files on slave machines.

Jmeter master-slave's issues

I am facing a problem in master-slave mode, we have a own jar to create some files after test, but when we run the test in master-slave mode, each slave will create one file, but we only need a summary file in master.
is there a way can instead of master-slave mode, let us use one jmeter server to test 2000~5000 threads? or change some setting to make the files only create in master?
JMeter simply executes the same test in the all the slaves. So, all the slaves will try to create the file as you had described if the test is designed to do that.
You could probably use a If Controller for a specific machine to do that.
If Controller ["master"=="${__machineName}"]
logic to create the file
I am facing a problem in master-slave mode, we have a own jar to
create some files after test, but when we run the test in master-slave
mode, each slave will create one file, but we only need a summary file
in master.
Since all the slaves can access the master machine, you can create a common shared file where in the slaves can make their entries based on the data that they have received. This can be done in 2 ways:
You can use a JSR223 sampler or a Beanshell sampler to write files to a common location thereby having just one master copy of the file
You can modify the JAR that you are utilizing right now to do the same work
is there a way can instead of master-slave mode, let us use one jmeter
server to test 2000~5000 threads? or change some setting to make the
files only create in master?
No. This is currently not a feasible option since the JAR that has been uploaded will be applicable on all the slave machines and hence will create multiple copies of the file.
Besides, if your master machine could generate 2000-5000 threads, I don't think you would have gone for a distributed load test in the first place.
Hope this helps!

JMeter Slave Reporting in Distributed Testing

I am performing distributed testing using JMeter and getting collective result of all slaves on Master JMeter GUI.
Please let me know if it is possible to capture separate reports for each slave?
For example, I have configured 2 Slaves + 1 Master JMeter. On executing the test, I would like capture separate report of each slave.
Thanks!!
You can get per-slave results as follows:
Add a Listener or your choice to Test Plan
Specify Filename in "Write results to file" section and choose what you need to store via "Configure" button
Once test finishes you'll have to collect result files from slave machines
Be aware that:
By default not everything can be stored, i.e. response data is being removed form successful samples
Listeners have their cost. See Greedy Listeners - Memory Leeches of Performance Testing to get an idea regarding impact and why it is better to have them off during load test.
For your case, don't use distributed testing, just launch 1 separate JMeter per slave machine.

How to get the combined result summary in master slave architecture using Jmeter

I have configured two slaves and one master machine.
After executing the performance test in Jmeter using the master slave architecture on master machine, I am getting the result log in slaves machines but I need a consolidated log summary. Please suggest me how can I get it?
Also, I need to know that how the load is distributed over the slave machines.
First question:
If you are talking about the logs (as in jmeter-server.log) there's no automatic way to collect them all. You could write a bash script or something similar to collect them from the servers.
Second question:
Every slave machine executes the full test plan completely on their own. That means that if you configure your test plan to include 10 threads, every slave will run with 10 threads giving you a total of 20.

In Jmeter distributed (remote) testing, is it possible to store results file on remote server?

I was using non-GUI option to perform a distributed load testing with Jmeter from master server A (Linux) to slave server B (Linux). Here's what I did:
./jmeter -n -t xxx.jmx -l results.log -R xxx.xxx.xxx.xxx
By default, Jmeter will write the results file to the master server A. I'm wondering if there's a way to directly store the results file on remote server B.
Thanks in advance!
The idea is they are dumb slaves, possibly without even the means to store data locally. The methods are invoked on a remote processor (RMI) and run from there, but all other resources are kept at the client side.
You could set up a CI server like Jenkins to run jmeter clients in a distributed manner. That will allow you more control over how tests are distributed. I have a jenkins set up that runs jmeter client on a master, with distributed tests to remote (jmeter) slaves, which are also jenkins slaves, tied to jobs that control the workspaces. This means they are never used by jenkins, but jenkins can get results files out of the workspace, aggregate them, and produce graphs using jmeter plugins CMDrunner.
The other possibility is to prefix your sample results with the unique part of the IP address of the current slave. This will allow you to display them aggregated or separately in your results viewer, and potentially write scripts to separate the results out and store them back on the slave they came from.
You could also use samba or similar to share the remote slave drive, and store the results directly there from the master.
If you have only one slave, any of these would be easy to implement, but get more complex as you add more slaves.
Look at sample sender alternative implementations.
Read the concerned paragraph here:
http://jmeter.apache.org/usermanual/remote-test.html
Natively it works like this, an option is not to use distributed mode and run many jmeter instances in non gui mode.
Try jmeter user mailing list and if you don't get an answer maybe open a bugzilla enhancement request.
If you want to build distributed load testing system you can try use Jagger

Resources