Page 1 of 1

General Question - Test Distribution

Posted: Mon May 11, 2009 6:14 pm
by atom
Hiya,

Not sure if anyone in Ranorex (or Community) can help me here.
We are developing GUI tests using Ranorex, and non-gui tests just as normal .exe's. But what we are lacking is a software package that will:

- distribute the tests over many machines
(with exception handling, re-routing, scheduling options etc..)
- collate results back to a central database
- provide management reports

Do you know of any such software?

It would be great to have an off the shelf product to do this, rather than developing some in-house solution

Regards

Posted: Mon May 11, 2009 10:07 pm
by Ciege
In my experience any test distribution that wasn't "roll your own" came from the developer of the test automation package and usually required the additional costs of test agent licenses. One in particualr I used in the past was from IBM/Rational where you could distribute your tests to other machines using a combination of test agents and Test Manger. It was clunky, hard to use and only worked with their products.

That being said, I know you don't want to do the development in house but that is what I would suggest. You can use the Ranorex run time licenses on various test machines to be able to run your automation code. You can use a free tool such as PSEXEC to distribute your tests to multiple machines. Then, by doing some basic development in your automation code you can create checkpoints to wait for things to happen on other machines. You can set up a centralized computer that each test client writes information to and the others read so that they can stay synchronized.

Not super simple but not beyond the capabilities of any automation expert worth their salt.

Good luck...

Posted: Tue May 12, 2009 4:47 pm
by Ciege
Also wanted to mention the use of virtual test machines. You can get free licenses for VMware or MS Virtual PC. From a remote location you can start/stop/reset the virtual PC to baseline configurations.

From you automation code you can control the states of these virtuals and copy or install your AUT across the wire and run tests. Then shutdown, reset and restart the virtuals to a baseline and do it all over again with your next iteration of your AUT.

This is very similar to what I do nightly with our builds. We automatically build, then reset the virtual PC, then copy the AUT installs to the PC, install and run the automated tests. I then have the build wait for the results of the automated tests and kick out an email with the status of the build and the status of the smoke tests that were executed.