Remote Execution Server gets hanged when Web Test hangs or is killed

[aplicaciones]

Hello,

My Scenario is the following one:

I have sent to the Runtime Execution Server one test list containing just one Web Test. This Web Test at some point gets hanged because presumably the web page does not connect to a pop up. It is quite probably something that I am not handling properly. But the issue I am facing right now is that the Test never finish nor expires. So, it would stay unfinished forever. So, I am killing the process (IE) manually.
But when I see the status of the Execution Servers it looks like the Test is still being run which is not true as I have manually closed IE.  So, I am not able to schedule nor run any more remote test on that Server.

I need a way to cancel that Specific Test List without  dropping the Scheduler Server DB because on other Runtimes I have several other Tests running properly. 
I have tried cancelled with Postman (following the Fiddler sample). It returns:
{
    "CancelledTestListId": [
        "0f7acba0-81c7-4de7-ad33-ee6831fc4773"
    ]
}
but I still see the Test in the status.
I have attached an image showing that shows the Runtime  "running" a Test even when I have killed any IE instance on that machine.

I hope my explanation is clear enough and that you could give me some hint on this.
Regards,
Sebastian

[Elena]

Hello Sebastian,

Thank you for sharing details on the faced issue. 

The described behavior is actually expected as the job is not properly terminated and the Scheduler did not get proper response. Though it is not quite convenient and could introduce further difficulties for not distributing results, etc. 

We are aware of that inconsistent behavior and we are working on improving the whole setup and communication between the components to provide a stable and satisfying user experience. Although that task is still in progress and there is no exact estimation when to be expected I would like to share we appreciate all feedback we get in regards the remote scheduling configuration. 

Also I would like to share an approach how to overwhelm that without dropping the database. Please follow these steps: 
1. Double check the Manage results view and delete any test lists without End TIme. 
2. Exit Test Studio and the Test Runners on each machine included in the remote setup. 
3. Restart the Telerik services on the machine which acts as a Scheduler. 
4. Type in the start menu Scheduling configuration on the machine which hosts the Scheduler. 
5. Click on Apply on each of the tabs. 
6. Start Test Studio and the Runners one by one. 
Even any previously scheduled tests are supposed to start after performing these actions. 

Another recommendation I could share is to build and construct the automation tests in a stable way to avoid such unexpected hangs. Then when the scripts are reliable enough you could run them on a remote machine. 

Thanks in advance for your cooperation. 

Regards,
Elena Tsvetkova
Progress Telerik