Forum Discussion

Fimbultyr's avatar
Fimbultyr
Contributor
7 years ago

Test Execute/Complete Service not starting reliably

Hi guys,

 

I am currently researching a phenomenon of my QA Environment regarding the Test Complete Services.

To be precise, TestComplete 12 Service does not start automatically, as it is intended and therefore my Network Suite throws a connection Error.

 

My Approach looks like this:

1. Run script to revert QA Machine VM through vSpeher CLI. (to a Clean state)

2. Run script to boot the targeted machine. (Server 2016 with Test Execute)

3. Execute Project that connects to this machine with connection Timeout of 10 Minutes.

 

Quite often i than get a timeout/connection error of this machine telling me:

WIN_2016_X64 (192.168.250.106): Invalid address. 

 

My next step would be pinging the targeted VM and of course it is answering properly.

Than i try to connect via RDC, and this works too.

 

I than look in the Services and see that Test Complete 12 Service is set to Automatic but is not running.

The Services that are listed in the dependencies are running as expected, but somehow the TC Service did not start.

 

In the Event Viewer I can find an Errors regarding the Test Complete Service:

1. A timeout was reached (30000 milliseconds) while waiting for the TestComplete 12 Service service to connect.

2. The TestComplete 12 Service service failed to start due to the following error: The service did not respond to the start or control request in a timely fashion.

 

I tried to reinstall Test Execute but this does not change the behavior.

Funny thing is, if I revert the Machine, let it start up and than login via VMware Workstation and normal windows login, the Service is running.

 

Maybe someone else has seen this issue before?

 

Cheers,

Constantin

  • Fimbultyr's avatar
    Fimbultyr
    7 years ago

    I kind of found a workaround for this, but i expect that Smart Bear may have a look in this issue as well.

     

    1. Ine the Networksuite/Jobs set the delay of the Test VMs to 10 Minutes Timeout.

    2. In Client OS, set the TestComplete Service from Automatic to Automatic(Delayed)

    3. In Client OS Registry, modify Timout of Services from 30.000ms to 60.000-120.000ms

     

    The Reg-Entry is:

    HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ServicesPipeTimeout

    If it is not existising, just create it as DWORD.

    Default Value is 30000, i have set it to 120000 and did not get an error after this.

     

    I will still monitor the VMs to see if this could be a longtime fix.

     

    Cheers,

    Constantin

  • tristaanogre's avatar
    tristaanogre
    Esteemed Contributor
    7 years ago

    This is a problem that I have on occasion on Windows 2012 R2 Server specifically after a reboot.  Once I get it started (several tries) it stays running until the next machine reboot.

  • Fimbultyr's avatar
    Fimbultyr
    Contributor
    7 years ago

    I kind of found a workaround for this, but i expect that Smart Bear may have a look in this issue as well.

     

    1. Ine the Networksuite/Jobs set the delay of the Test VMs to 10 Minutes Timeout.

    2. In Client OS, set the TestComplete Service from Automatic to Automatic(Delayed)

    3. In Client OS Registry, modify Timout of Services from 30.000ms to 60.000-120.000ms

     

    The Reg-Entry is:

    HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\ServicesPipeTimeout

    If it is not existising, just create it as DWORD.

    Default Value is 30000, i have set it to 120000 and did not get an error after this.

     

    I will still monitor the VMs to see if this could be a longtime fix.

     

    Cheers,

    Constantin

    • temich6059's avatar
      temich6059
      New Contributor
      7 years ago

      I am also getting this same type of error as well.

  • Fimbultyr's avatar
    Fimbultyr
    Contributor
    7 years ago

    There are 2 additional Steps you can try.

     

    1.)You could as well try to improve the performance of your System.

     

    I benchmarked my VM Environment and found some issues with the ESXI HDDs.

     

    If I let my VMs run on the normal SATA-HDDs on ESXI, I encountered boot times of around 9 Minutes for Windows 10 VMs.

    After moving them to SATA-SDD of the ESXI the VMs now boot up in 30 Seconds.

     

    This solved a lot of issues for me.

     

    2.) I activated automatic sign in to the OS.

    So basically after booting the test VMs, they now log in automatically to the Administrator account.

    This is of course a security issue and should only be done in controlled and isolated QA Environments.

     

    I had the impression that the RDC Connection from Testcomplete somehow caused an Issue with the bootup of the Service.

    If you log in to the account manually before connecting via RDC the Service did run more reliably.