Community Tip - If community subscription notifications are filling up your inbox you can set up a daily digest and get all your notifications in a single email. X
I am trying to go through 'Build and Deploy an Analytics Model' guide. When I look at Analytics Manager and click on Connected Clients I see the following message - No data available for Connected Clients
The guide asks to restart the server as administrator which I can't do with a hosted trial.
Appreciate any help in getting this working
Solved! Go to Solution.
Hi @cnuga
This is where the issue is.
The guide specify you need to use Analytics Trial Edition from https://developer.thingworx.com/resources/downloads/analytics-trial-edition#
not ThingWorx Trial Edition.
The Analytics Trial Edition include ThingWorx Foundation and the Analytics microservices.
It does not however include the industrial connectivity.
However if you want to use ThingPredictor, you will need the Analytics Trial Edition, not the ThingWorx Trial Edition.
It is also important to make sure to have AnalyticsTrialEdition-8.2.0-r3-windows-installer.exe as older version (r2 for example) won't work.
Kind regards
Christophe
Hi,
I believe creating an Analysis Provider will resolve the error that you are encountering. Could you follow the steps at this link and see if that resolves the issue?
Warm Regards,
John
Hi @cnuga
The guide https://developer.thingworx.com/resources/guides/analytics-experience requires to have a locally installed Analytics Trial Edition (ATE)
ThingPredictor is not implemented on the hosted instance.
See The Environment chapter at the bottom of https://developer.thingworx.com/resources/guides/analytics-experience
Kind regards
Christophe
I have installed Thingworx trial edition on my computer, but I am running into the same issue. Am I missing something? I did all TW.AnalysisServices.ThingPredictorConnector in Connector (Required) field.
I tried the solution in the guide
docker-machine ssh DockerMachinePTC docker restart predictionms
but get the following error
docker-machine : Docker machine "DockerMachinePTC" does not exist. Use "docker-machine ls" to list machines. Use "docker-machine create" to add a
new one.
At line:1 char:1
+ docker-machine ssh DockerMachinePTC docker restart predictionms
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : NotSpecified: (Docker machine ... add a new one.:String) [], RemoteException
+ FullyQualifiedErrorId : NativeCommandError
I am having issues with the Thingworx Analytics Server Installation. I have tried the guide "Build a Predictive Analytics Model' and get stuck at creating signals. I get the following error - NotFoundException: Unable to retrieve data URI for the dataset [6096e1bf-d8cf-49cf-9da3-255968a55acd] with message
The dataset exists, I've checked.
Also, when I try the guide "Build and Deploy an Analytics Model" - I check for 'Connected Clients' in analytics manager. I don't see a green square i.e., connection between ThingPredictor and Analytics Manager is not established.
I suspect both these problems are occuring because the Analytics Servers is not configured properly. I have attached the properties file.
Hi @cnuga
Could you specify which OS version did you install on ?
What docker deployment did you install: Docker for Windows or Docker Toolbox ?
This will help to know what commands will be appropriate to troubleshoot.
Thanks
Christophe
Windows 10 Pro and Docker Community Edition Version 18.06.0-ce-win72 (19098)
ok so this would explain the error on DockerMachinePTC.
DockerMachinePTC would only exist if using docker toolbox on windows 7.
Can you execute docker restart predictionms, and check the Connected Client UI.
If you still see no connected client, could you post the output of :
- docker ps -a
- docker logs predictionms > prediction.log ... and send the prediction.log created.
Kind regards
Christophe
It still doesn't show a connection on the Connected Client. Attached are the two logs, one for docker ps -a (Docker.log) and Prediction.log
Hi @cnuga
Thank you for this. I see something that is likely going to create problem.
Before going forward though I would like to get more data if possible.
could you check under %USERPROFILE% if you have a twx-analytics folder.
If you do, could you zip it and post it here ?
note to save space you can remove the db files from the zip.
Thank you
Kind rgeards
Christophe
Thank you for this.
It appears that you have got another application running on port 80 on this machine.
This breaks the internal communication between the different microservices as nginx is not running on the expected 80 port.
I do not know if this is the root cause of the issue you have with thingpredictor but this is certainly the reason for the issue with signal.
Could you please try the following:
- Uninstall ThingWorx Analytics Trial Edition (ATE)
- Reboot the machine
- Ensure that port 80, 8080, 8900, 9000 9100 9200 9300 9400 9500 9600 9700 are free as they are needed by the ATE.
- install ATE
Try again signals and Thingpredictor.
Note, do not change the configuration files. Some of the ip addresses are meant to be from the internal docker network and not the actual machine ip, changing this will break the configuration.
Thank you
Kind regards
Christophe
Thingworx is running on 8080. A windows service ntoskrnl is running on 80. I have changed to 81. Other ports are being used by Docker. Restarted Thingworx Foundation and Analytics Server but signals is not working - get ConnectException: Connection refused (Connection refused).
Should I reinstall Analytics Server?
yes you do need to reinstall as the nginx container needs to be recreated by the installer.
You do need to follow the steps I mentioned earlier:
uninstall ATE
reboot
make sure ports are free
reinstall ATE
note you mentioned ThingWorx is running on port 8080. I just want to clarify that there is a ThingWorx installer with ATE and it does need to run on port 8080. So you cannot have another ThingWorx running on the same machine on port 8080. From the docker log this part seemed ok, but I wanted to make it clear .
Kind regards
Christophe
The signals part in Analytics Builder works fine but the ThingPredictor connector is still showing no connected clients.
Hi @cmorfin how can I run it on the hosted server, in that case, does it mean the analytics cant run from the hosted server ?
Greetings,
At present, the 30-day hosted DevPortal evaluation server has both Foundation 8.2 and Analytics. It does *NOT* have ThingPredictor, though.
As such, both the Analytics Builder and Manager guides (as well as the combined "Build and Deploy" guide) recommend running through them vs the Analytics Trial Edition (ATE) 120-day downloadable trial.
ATE *DOES* include ThingPredictor, so there's no issue with publishing your completed model and running Analysis Events against it.
Thank you,
Jason
The predictor connection is not working in trial server installation? Is there a way to fix it?
If you are having issues with ThingPredictor connecting to ThingWorx Analytics, please refer to this page of the "Operationalize" guide on Analytics Builder:
https://developer.thingworx.com/resources/guides/operationalize-analytics-model/analysis-provider
Specifically, the bottom section copied here:
ThingPredictor is installed and configured to point to Analytics Manager automatically as part of the installation of the Analytics Trial Edition (ATE).
At maximum, only the last 48 hours are shown, so connectivity with ThingPredictor can be difficult to confirm if it has been longer than that time period. In addition, ThingPredictor can timeout if you wait too long to configure a ThingPredictor Analysis Provider after installing ATE.
To force ThingPredictor to restart, enter the following command from a command line interface as an ADMINISTRATOR:
docker-machine ssh DockerMachinePTC docker restart predictionms
Thank you,
Jason
Hey, @jwyatt I now have a new doubt about the whole usage of the application I'm now developing all the solutions on the hosted server now with the hosted server not having the ThingPredictor. should I change my development to local servers?
will the premium version have the thing predictor or is there any way we can add the predictor to the hosted version
The DevPortal *HIGHLY* recommends that all individual-application / PoC development occur on one of the 120-day downloadable trials.
We recommend that the 30-day evaluation server be used solely to run through the various guides to get a feel for the platform and its capabilities.
As soon as you're developing your own application, please do make use of one of the 120-day downloadable trials.
In particular, the 30-day evaluation server is auto-deleted at midnight on the 30th day of your trial. We have no way to recover lost work.
Thank you,
Jason
To get my Thingworx Predictor connect to work, do I need to install Thingworx Agent as shown in this article - https://www.ptc.com/en/support/article?n=CS258412
Hi @cnuga
No you do not need to install anything. Everything is included in the trial edition.
As mentionned in the guide they can be a timing issue so it is important to restart the predictionms. So I woudl suggest the following:
- execute docker restart predictionms
- wait 1 min
- check the connected client UI - take a screenshot of it to post here too
- irrespective of the result in connected client, perform the model publish, to check if it works or not.
If it does not work, upload the following data (similar to previously but we need to see the current one after the reinstall):
- output of docker ps -a
- zip of %USERPROFILE%/twx-analytics folder
- execute docker cp twxfoundation/ThingworxStorage/logs .
Note there is a . character at the end of the above command.
zip the logs folder, created by docker cp command, in the current directory
Thank you
Kind regards
Christophe
Attached please find the screenshot with error message.
The docker ps -a log file
The Zip file.
I couldn't get docker cp twxfoundation/ThingworxStorage/logs . to run in Windows Powershell
I am in PS C:\Users\cnuga> and when I type the command I get docker : must specify at least one container source
At line:1 char:1
+ docker cp twxfoundation/ThingworxStorage/logs .
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : NotSpecified: (must specify at...ontainer source:String) [], RemoteException
+ FullyQualifiedErrorId : NativeCommandError
Hi @cnuga
Apologies the command is docker cp twxfoundation:/ThingworxStorage/logs .
Thank you
Christophe
Hi @cnuga
I think I find what is wrong.
could you tell me :
- the exact name of the installer you used to install ThingWorx Analytics Trial Edition
- when did you download it
I just realize something is missing in the properties file, so you are not using a version that is configured properly for ThingPredictor.
Once I have the answer to the above question we can see how to proceed.
Thank you
Kind regards
Christophe
The name of the installer is ThingWorx-Trial-Edition-8-2
Downloaded on 21st August, 2018.
I reinstalled both Foundation and Analytics yesterday
Hi @cnuga
This is where the issue is.
The guide specify you need to use Analytics Trial Edition from https://developer.thingworx.com/resources/downloads/analytics-trial-edition#
not ThingWorx Trial Edition.
The Analytics Trial Edition include ThingWorx Foundation and the Analytics microservices.
It does not however include the industrial connectivity.
However if you want to use ThingPredictor, you will need the Analytics Trial Edition, not the ThingWorx Trial Edition.
It is also important to make sure to have AnalyticsTrialEdition-8.2.0-r3-windows-installer.exe as older version (r2 for example) won't work.
Kind regards
Christophe