[ptt-users] Some more TestMaker5 questions

[Frank Cohen]

Hi Sameer:

Thanks for using PushToTest TestMaker. I am very glad to see it is  
working at CA. Here are some thoughts on your questions:

> Question# 1: When I try to run the TestMaker5 from the command  
> line, its spits out the message: "The node localhost is invalid".  
> However, when I have TestMaker GUI up and running and then I try to  
> run a scenario via command line, things work just fine. I am using  
> the following command line:
>
>
>
> TestMaker.bat -commandline -t Tests/Trade6LoadTest.xml
>
>
>
> I am assuming that somehow this command line itself is not able to  
> start the TestNode and therefore complains when run without the  
> TestMaker GUI up and running. Is this a known problem or I am  
> missing some parameter in the command line? It seems odd to have  
> this kind of dependency on the GUI.
>

PushToTest TestMaker starts a local TestNode as a convenience when  
authoring tests in the TestMaker application. Please launch the  
TestNode from the command-line too. You can do this today with:

TestMaker_home/TestNetwork/TestNode/startup.bat (for Windows, and .sh  
for Unix/Linux/Mac)

(By the way, the same is true about the local Resource Monitor -  
PTTMonitor)


>  ·         Question# 2: TestMaker logs data for each transaction  
> after every run (in xml format). Depending on the transactions  
> throughput and test durations, this can be a huge xml file and TM  
> itself can take up significant time and resources to generate that  
> log (at the end of the run). Is there a way to shutoff this logging  
> if that raw data is not required?
>

There are controls over the logging provided in the TestScenario.  
(http://docs.pushtotest.com/docs/reference.html) There is currently  
no way to eliminate the transaction logging. I am open to your ideas  
to make TestMaker better and easier to use.

We are currently working on an Enhanced Results Analysis engine that  
gives you much more control over what is reported. Details are found  
at http://bugs.pushtotest.com/ticket/188 and I blogged about it just  
yesterday at http://www.pushtotest.com/thecohenblog.



> ·         Question# 3: This is continuation of Question# 2. During  
> any given test run, how is testmaker5 internally managing this per  
> transaction data in its internal data structures? For high volume  
> tests spanning over many hrs, are we supposed to run into JVM heap  
> out-of-space kind of problems with the way this is implemented  
> right now?
>

TestMaker uses a snapshot mechanism to move the transaction data from  
the TestNodes to the TestMaker console. It is possible to run out of  
memory but we have not seen that happen yet (and we have run multi- 
day tests.)

One option you may find useful is logging to a relational database.  
This is documented here: http://bugs.pushtotest.com/ticket/62 and  
results analysis in general is at:  http://docs.pushtotest.com/docs/ 
analyzingresults.html


Hope this helps!

-Frank


>
>
>
>
> Thanks in advance,
>
> -Sameer
>
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.pushtotest.com
> http://lists.pushtotest.com/mailman/listinfo/users



--
Frank Cohen, PushToTest, http://www.PushToTest.com, phone 408 871 0122
TestMaker open-source test automation