Forum Discussion

pvidal's avatar
pvidal
Contributor
15 years ago

Real-time status for running test.

Hello,



We have a test project with about 200 test items that takes several hours to complete. When the test is running we see the tested application being exercised but we don't know which of the tests is being executed or if a test has failed until the end of the execution. Is there a way to have some kind of "live" status so we know if a test already has failed and which test is currently executing?.



Thank you,

Patricio.

6 Replies

  • hlalumiere's avatar
    hlalumiere
    Regular Contributor
    For our scale of execution we found the included tools for distributed testing management somewhat inadequate. Here is a pic of our main management dashboard, that interfaces to a SQL database, and of our distributed client, that configures test parameters prior to running each test, fetches the next test according to priority, sorts out associated development tickets being fixed by the dev team, updates the test suite from out SVN repo, runs TC through COM and updates the database with results, restart VMs that are failing repeatedly, serves as a watchdog, etc... Our tests run continuously, 24/7, the database is updated with results in real time, we fix tests all day long, have the possibility to throw a particular test at the top of the heap to get results on demand, etc... We also have another tool that runs on a management VM that reverts and restarts offline VMs automagically. Since we have upwards of 45 VMs total, we manage CPU and memory usage through vSphere directly.



    Intentionally blurred for your viewing pleasure...
  • Using TestComplete 9.10 I created a Userform with a status bar that is updated as the tests are performed.



    It is event driven, and posts the same messages that go into the log, to the the appropriate window (I only have 2 for now since I only care about errors). As well, I keep a running count of the events as temporary project suite variables. The entire thing is updated all at once using a single function to stick it all together.



    The tests running out of overall number of tests, is precalculated using a separate function and the status bar is just a reflection of that (#of test /100).



    The CPU and Memory meters are event driven as well which is good enough for now...9.20 apparently has some additional things to tap into that could make those a little nicer :)
  • To know anything during the test run, I think you ll have to use TestComplete Indication and have all the data that  you would like to see pushed into the indicator. Please have a look into the help topic "Using TestComplete Indicator"



    Thankyou








  • cjayesh's avatar
    cjayesh
    Occasional Contributor
    Or if you are sitting away from the testmachine, you could possibly call a routine at the end of every test to email a notifier to you about the progress, though in your case it means you will have 200 emails per test run.
  • Hi Patricio,


    I have added your request to our DB as a suggestion. Thank you for your post.


    In addition to Jayesh's and Ahmed's posts, you can work with TestComplete via COM (see the 'Working With TestComplete via COM' help topic), so you can create a utility that will obtain and show the name of the currently running test item.


  • Hello,



    Thank you all for your replies. I am using Ahmed suggestion and pushing information to the Indicator. Jayesh suggestion is interesting and I may use to notify final status.



    Allen, thank you for submitting my request. The COM interface seems to be the way to get all the information I need. I am sure that many users would find use for that utility so maybe AutomatedQA could develop a "TestComplete Live Status" tool :-).



    In the meantime I am also logging the beginning and end of each test case, and any failures to a text file and using the free BareTail tool (http://www.baremetalsoft.com/baretail/) to monitor the file. The good thing is that with BareTail I can color specific tags so I can show errors in red, and also I can monitor tests running in other computers (or Virtual Machines) by sharing the folder where the log is written.



    Thank you,

    Patricio