Hey Guys,
I am trying to setup our Test Studio system as indicated by the attached network diagram. After opening the correct ports and various configuration checking I am now able to connect to the Scheduling Server from Test Studio on the Authoring Machine. In Test Studio on the Authoring Machine I am able to see that the Scheduling Server's Execution Service and the Execution Server's Execution Service are both connected to the Scheduling Server's Scheduling Service. I try to schedule or run a list remotely and the program hangs waiting to Update Latest Project Files. After about 5 minutes it errors.
The execution service logs are empty and the scheduling service log has the following repeated hundreds of times:
[03/24 08:43:49,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobBroker.ctor() : Job runner refused this job!
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobRunner.RunJobAsync() : Unable to load test list, aborting run.
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobRunner.RunJobAsync() : Use source control server: NO
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobRunner.RunJobAsync() : Project source location: C:\svn\TestLibrary\UI Automation\SEWeb
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobRunner.RunJobAsync() : Check if test list is available at project location. Consider cleaning Job storage!
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobRunner.RunJobAsync() : Job storage location: C:\Windows\system32\config\systemprofile\AppData\Roaming\Telerik\TestStudio\Scheduler\ScheduleStorage
[03/24 08:44:19,Telerik.TestStudio.ExecutionManagerService.exe(2040:24),Execution] JobBroker.ctor() : Job runner refused this job!
I have all three computers sync with SVN repositories in the exact same file structure as indicated by the attached picture.
I have read other posts and documents and have tried the following:
- Changing the Scheduling and Storage Service to use an administrator account
- Attempting to run the test on the Scheduling Server's Execution Service
- Turning off all 3 Computers' Firewalls
- Reinstalling the Scheduling, Storage, and Execution Services on the Scheduling Server
As another note, I was able to get a test to run remotely previously while on a Trial of the software on different VMs.
Please help me see what I am missing. I have read the following posts to no avail at solving my current issue:
unhandled-error-when-trying-to-run-list-remotely
test-list-not-running-on-execution-machine
Thank you in advance for any assistance,
Kelly
16 Answers, 1 is accepted
Rank 1
Similar things happened to us when we began with our setup, somewhat similar to yours. Our setup involves the following...1 computer holding SVN repos, running the telerik scheduling server, and running the telerik storage server. Another computer used for our teamcity build agent and telerik execution server. And then whatever other computer is being used to develop tests on. It looks like you have covered the basic suggestions I began with. Nothing jumps out to me in particular but I can do my best to give you some advice.
First things first: restart the telerik storage service and telerik scheduling service from the scheduling servers task manager or services menu (gotten to from the services button in task manager). Give running remotely a shot again.
1) We started out by getting our configuration running with the scheduling server & execution server on the same computer (I think that is what you were getting at in bullet point 2). Without any teamcity building involved, simply running test lists remotely from the development machine. Also, I did this with a completely new project that just did something really simple (like clicking on something) with the application we test against.
To address: Consider cleaning Job storage!
2) Have you ever dropped your old database? This seems to help remote execution issues significantly for us but beware, you will lose all your existing results in the database you drop. There is probobly a way to back it up but never looked into it. This can be done through the use a program by the name of Robomongo, a GUI interface to mongo databases (test studio storage, unless your using your own database or SQL this does not apply). Once you download it on the machine with scheduling & storage you can use the following settings to connect to the Test Studio storage database if everything you used to set up were the defaults. Connection Name: Test Studio Storage Database. Address: localhost:27017. Once in there, if set up correctly you should see your Test Studio Storage Database on the left hand side of the screen. To drop your database, right click TSStorageData and select Drop Database. The database will be recreated when you run a test list remotely again. I hope this one fixes your issue right here! It has been useful on numerous occasions, especially if you are running remotely and you get an error saying a test or test class type can't be found.
3) After re-installing everything, did you delete the C:\mongoDB file (or something similar to that) that the storage part of the installation left behind? This is something I did, whether it helped or not I am not exactly sure. I'm a bit superstitious when it comes to this kind of stuff.
4) This was suggested to me but didn't apply: if you have anti-virus turn it off.
5) I re-call that one specific issue that took me a bit to figure out. The machine I was installing my scheduling & storage servers on had test studio installed on it a while ago. One of the ports test studio needed was being used by an older telerik sql service causing connection issues between my scheduling and execution server.
As far as debugging ports goes... You can figure out which applications are using which ports on your scheduling server by using a dos command to display the PIDs of the applications using the ports on the machine. Then you can use your task manager to map the PID to an application name if you want to be confident the proper ports are being used by telerik services.
Now I'm just gunna spit ball a bit.
6) If you are testing against a WPF application...all of the projects referenced DLL's must be either located in the given projects root directory or in addition referenced from there as well.
7) If you can view your scheduling servers log from your development machine then at least you know that if its a communication issue, it isn't located there.
Also there are a few logs that one can look into. Here are some places to check if you haven't already.
- Development machine Test Studio Log (C:\ProgramFiles (x86)\Telerik\Test Studio\Logs).
- Scheduling server Test Studio Log (same location).
- Execution server Test Studio Log (same location).
Let us know what you learn!
Rank 2
Hey Chun,
Thank you so much for your thorough reply. I have tried figuring this out in so many ways. First thing I would like to mention is that we don't have any existing test database nor do we have any tests other than go to a home page and click login.
I have gone as far as to spin up to brand new machines, install the entire Test Studio, Scheduling Service, and Storage Service, and scheduling tests, and trying to run them remotely. I can schedule the tests locally, but I cannot get anything to run/schedule from the second machine. I have tried both ways. These machines are both fresh installations of Windows Server 2012 R2.
I have done the above steps twice now. Once with Windows 10 to see if that was an issue.
I cannot get two machines to talk together other than connection, status, and reading the remote log. When it comes to "Updating Latest Project Files" it fails every time in many different machines.
Thank you for any further help :-)
Kelly
Rank 2
Howdy,
At this point I would like to mention that I have taken any source control or repository out of the equation. I have a test that navigates to a page and nothing else. I have a test list that runs that test. I try to run that test remotely and I receive the same error. This is performed on two brand new Windows Server 2012 R2 VMs in Azure that have their firewalls turned off.
There is no SVN, no TFS, nothing but Test Studio Web and Desktop on one and Scheduling, Storage, and Execution on the other. The same error occurs after attempting to update latest project files.
I have watched the training video and have seen these exact steps work correctly without TFS or any source control. I am unsure why everything appears to be connecting and I can see the Execution Servers in the status list, but I cannot send any tests to them.
Thank you in advance for your help.
Thank you for the detailed information and the public thread. I share the update of your support ticket here as well. And as it duplicates this thread here I would recommend to close the support ticket to avoid double work.
As you mentioned the two machines are Windows Server 2012 R2 VMs, please read through this article - the IE enhanced security configuration should be turned off. Another important thing for scheduling and remote execution are the ports used by Test Studio to be opened. Please ensure that the following incoming and outgoing ports are unblocked on the appropriate machines:
Scheduling Server: 8009
Storage Server: 8492
Execution Machine(s): 55555, 8030-8039 (and sometimes a range of ports from 8040 upward).
Please let me know what you mean by "attempting to update latest project files" - 'Get latest' feature is enabled only when a project is connected to a source control system. Please elaborate further details on that topic.
If you continue experiencing troubles with the scheduling configuration, please let me know a suitable time frame when you could attend a web meeting and your time zone so that I could arrange one in appropriate for both sides time.
Thank you for the cooperation and hope to hear from you soon.
Regards,
Elena Tsvetkova
Telerik
Rank 2
Hey Elena,
On the VMs in question IE Enhanced Security is turned off. Also, I had previously configured those ports to be open, but after having further trouble, I turned off all firewalls on all machines in question.
I do not have the project connected to source control but I still get that message. This is where I think the problem may lie.
Thanks for your help!
Kelly
PS - I am available from 7:00a.m. - 4:00p.m. Central Time CDT (12:00 - 21:00 UTC).
I provide you the meeting details here as well:
1. Please join my meeting.
https://global.gotomeeting.com/join/944574741
2. Use your microphone and speakers (VoIP) - a headset is recommended. Or, call in using your telephone.
Dial +1 (571) 317-3122
Access Code: 944-574-741
Audio PIN: Shown after joining the meeting
Meeting ID: 944-574-741
GoToMeeting®
Online Meetings Made Easy®
Not at your computer? Click the link to join this meeting from your iPhone®, iPad®, Android® or Windows Phone® device via the GoToMeeting app.
Please ensure that User Account Control (UAC) is turned off before we start the meeting in order to be able to grant control on the affected machine. Here is an article how to turn it off.
Hope to hear from you soon.
Regards,
Elena Tsvetkova
Telerik
Rank 2
Hey Testers out there,
I would like to recap our meeting with Elena and team. First of all thank you so much to Test Studio and their helpful team in getting all my servers working correctly with Test Studio.
In a word it all boiled down to fully qualified domain names as the host name for the scheduling service. In the configuration of the scheduling server we needed to change the server name to the FQDN for the hostname. Then on the authoring machines we needed to connect to that FQDN for the scheduling service.
This solved all connection issues and I have now been able to set up multiple executions servers and distribute test lists to them.
Thanks again :-) !
Thank you for your time last week and for your patience while we figure out the exact cause of the connection issue. I hope it will all worked smoothly in the future. However in case of any inconsistent behavior do not hesitate to contact Telerik Test Studio support team.
Wish you great week ahead!
Regards,
Elena Tsvetkova
Telerik
Rank 1
Rank 2
I believe this is the location you are looking for. Be patient, changes to this field take up to a minute to actually show you that it changed.
Rank 1
Rank 2
Rank 2
I believe that is set in Windows Computer Properties. You might need to be part of a Domain for it to have a proper FQDN.
I attached where to find that on Windows Computer Properties.
Rank 2
I believe that is set in Windows. You might have to be part of a Domain for a FQDN to be retrieved.
I have attached where to find this in Windows Computer Properties.
Rank 1
Yeah, under normal circumstances that makes sense.. I'm running a VM in Azure. The machine name ends up being "MachineName" where my FQDN is really "MachineName.westus.cloudapp.azure.com" But I can't set this as the machine or domain name. There isn't a westus.cloudapp.azure.com DNS server.
I'm working with support on this and will post my conclusions here.
Thanks
I assume your issue was successfully resolved after a meeting was conducted. It would be great if you leave a short summary of the actions that were performed to be able to use the cloud VM as an execution machine. Thank you in advance!
Regards,
Elena Tsvetkova
Telerik by Progress