jmeter is not running script in the order it is written - jmeter

In jmeter, I'm using nested loop controllers, along with some JSR223 postprocessors inside each loop
Here's the order the script is written:
Loop Controller
Loop Controller
http request
Endloop
JSR223 postprocessor
Endloop
I want the script to first run the http request (multiple times), then after that looping is complete, run the JSR223 postprocessor, then repeat all that.
Instead, what is happening is it enters into the first Loop Controller, then runs the JSR223 postprocessor, then runs the second nested Loop Controller.
Why? How do I get it to run the script in the order in which it's written from top to bottom and nested?

Change from using postprocessor , which is executed for every request in scope, to sampler which executed once
Some elements in the test trees are strictly hierarchical (Listeners, Config Elements, Post-Processors, Pre-Processors, Assertions, Timers), and some are primarily ordered (controllers, samplers). 
Another option is to add it under Sampler which will execute it omce after sampler

JSR223 PostProcessor is being executed after each Sampler in its Scope, in your case after each iteration of the HTTP Request sampler.
If you want to run it only once - either put Flow Control Action sampler at the place of the JSR223 PostProcessor and make the JSR223 PostProcessor a child of the Flow Control Action sampler.
Otherwise you can use JSR223 Sampler instead of the JSR223 PostProcessor, if you don't want the JSR223 Sampler to appear in test results - put SampleResult.setIgnore() function somewhere in your script.

Related

How to run a JSR223 PreProcessor script once per loop

I currently have a loop that executes a bunch of queries via JDBC request samplers. They should all share a random ID that changes on every loop.
I tried using the beanshell script and JSR223 PreProcessor. But the PreProcessor gets executed before every single JDBC reuqest sampler, not once per loop. I feel like there is an obvious fix to my problem that I am missing.
I also tried putting the JSR223 script into an "Only Once Controller". But then the random variable I inject with vars.put() is not visible to the JDBC sampler. Also, as far as i understand the Only Once Controller, it would only execute on the first loop iteration. Which is not what I want.
JSR223 PreProcessor obeys JMeter Scoping Rules so if you put it to be a child of i.e. q_insert1 sampler - it will be executed only once per iteration.
I would also recommend re-considering language selection, since JMeter 3.1 it's recommended to use Groovy for scripting

How to filter out last sampler in While controller in JMeter?

My test plan is as follows:
Thread group
\_..other items
\_While controller (with blank condition)
\_Web socket single read sampler
\_..other items
My problem is, because I use blank condition, the While controller executes till the sampler errors out. So my question is, is there a way to filter out this last sampler? There are no other conditions I can use in the While controller as the number of times the child sampler has to be executed is not constant.
Thanks!
Add JSR223 PostProcessor as a child of the WebSocket Single Read Sampler
Put the following code into "Script" area:
if (!prev.isSuccessful()) {
prev.setIgnore()
}
In the above code prev stands for the parent SampleResult and the setIgnore() function tells JMeter to not to collect the result for the given Sampler.
More information on this and other JMeter API shortcuts: Top 8 JMeter Java Classes You Should Be Using with Groovy

Difference between a JSR223 Sampler and JSR223 PreProcessor

I'm new to Jmeter and vigorously learning.
I wanted to know how JSR223 Sampler and JSR223 PreProcessor are different, can a sampler be used the same way as the preprocessor?
General difference is that PreProcessor won't be executed unless it has a sampler in its scope which it will be triggered (per sampler)
Pre-Processor element is defined to alter the settings of Samplers in their scope. It will always execute before the actual sampler request.
Specific difference is that JSR223 PreProcessor doesn't have SampleResult available, so for example you can't execute the following example:
SampleResult.setStopTest(true);
The SampleResult ResponseData is set from the return value of the script. If the script returns null, it can set the response directly, by using the method SampleResult.setResponseData(data), where data is either a String or a byte array. The data type defaults to "text", but can be set to binary by using the method SampleResult.setDataType(SampleResult.BINARY).
The SampleResult variable gives the script full access to all the fields and methods in the SampleResult. For example, the script has access to the methods setStopThread(boolean) and setStopTest(boolean).
JSR223 Sampler is a Sampler, so it will generate a SampleResult which will appear in the Test Results (unless you call SampleResult.setIgnore() method)
JSR223 PreProcessor cannot be executed per se, you need to connect it to one (or many) Samplers according to JMeter Scoping Rules so it will be executed before one (or many) samplers. PreProcessors execution time is not reflected in the test results (unless you use a Transaction Controller configured to include it)
Both can run an arbitrary code (it's recommended to stick to Groovy) so which one is to use mostly depends on your use case, if you need to conduct the load and measure the time - go for the Sampler, if you need to set up some data - go for the PreProcessor, etc.

jmeter - How to ignore selected samplers from getting measured in the *jtl file

My test configuration :
Loop Controller
--> Beanshell Sampler
vars.put("test", "${__CSVRead(*test,0)}");
Add it to an existing array retrieved from vars.getObject
In the above scenario, I am constructing my request payload dynamically in a loop controller. I had to put the CSVRead function in a separate Beanshell sampler under the loop controller since "${__CSVRead(*test,0)}" was reading the sample line if I use it within a for loop inside the beanshell sampler (interpreted mode).
While the above configuration meets my requirement, my *.jtl files are growing in size even for a 30 minute load test since the BeanShell sampler is getting measured all the time. While I am able to filter the required data by using the FilterResults tool, I want to know how to avoid this during the execution itself like the TestActionSampler
Use one of the following Test Elements instead:
BeanShell PreProcessor
BeanShell PostProcessor
Beanshell Timer
By default Timers and Pre/Post Processors execution time is not included into parent sampler elapsed time (unless you use Transaction Controller explicitly configured to do so), using this approach you will be able to exclude the time, required for constructing payload from test results.
I resolved it by using the following configuration.
Loop Controller
--> Test Action Sampler
--> Beanshell timer returning 0 at the end
vars.put("test", "${__CSVRead(*test,0)}");
Add it to an existing array retrieved from vars.getObject
return 0;

How to generate multiple jmeter sampler results from a BSF Sampler?

I have a BSF sampler that runs a groovy script.
I want to create multiple sample results out of this single script in jmeter.
Is it possible?
...my groovy script will run a junit suite and i want to report each junit test separately. Or my groovy script will make 100 http get and i want to display 100 sampler results.
To achieve this you have to organize loop - e.g. put your BSF Sampler under any suitable logic controller: Loop Controller), While Controller, ForEach Controller, or use possibilities of Thread Group (Loop Count, Number of Threads fields).
This scenario will re-use single instance of your BSF Sampler for each new iteration / thread and produce separate sampler result for each execution.
As well you possibly have to provide different entry data / params to BSF Sampler on each iteration - to parametrize it. In this case you can look at least into CSV Dataset Config (to read from file) or any of postprocessors/extractors (go get data from responses).
Suppose you also look into JSR223 Sampler to use along with groovy scripts for better performance.

Resources