How to stop an analytics running (profile) job using the Builder/Server/Api ?

[BernardoAlb]

Hi all,

So this as happened three times now: We run a profile job and it is eternally calculating and not succeeding in retrieving results.

I would like to ask if there is a way to stop an analytics running job ? (apparently a running profile job cannot be deleted in the builder).

Also I would like to know if this has happened to any one of you ?

I suspect this might be a bug, because in the other 2 times this happened eventually it failed (after 1 day). And the error seemed to be a Java exception for an impossible convertion between Integer and String, which is explained by a field of data type: Integer and opType: Categorical. Changing the data type to String solved this problem. 

This impeded our (and anyone's) work since no new profiles can be processed until the last one finishes.

Thank you for all your time,

Regards, 

Bernardo

[nsampat]

Bernardo,

Thank you for posting your question to the Developer Community.

I have a few questions, so that I may better address your question:

• What version of Analytics are you running?
  • Are you using the Analytics Trial Edition or the ThingWorx Trial Edition?
• What operating System is the Analytics Application installed to?
• Is this a native install, or did you use the docker deployment offering?
• Does this issue with Profiles happen with all jobs, or just with one particular dataset?

You are correct that in the current offering of ThingWorx Analytics, there is not a feature to stop or terminate a runaway job(s) and delete them.

Based on your answers to my clarifying questions above, we may have a work around for you to remove the runaway jobs.

There is a feature Enhancement request to enable a user to stop and terminate a hung job, and allow them to delete it.

I look forward to your response.

Edit: Bernardo, I also see that you have a case with one of my colleagues for this same issue. The triage and information I requested are also requested in that case. If you could kindly provide that information to my colleague via the case, that would ensure that engineer can assist you in your issue.

Regards,

Neel

[BernardoAlb]

Neel,

Thank you for your time and response.

We are using Thingworkx version 8.1 and Analytics version 8.1.040000

Analytics is installed as docker in a Linus OS.
(Docker Version:)

Client:
Version: 18.03.1-ce
API version: 1.37
OS/Arch: linux/amd64
Experimental: false
Orchestrator: swarm

Server:
Engine:
Version: 18.03.1-ce
API version: 1.37 (minimum version 1.12)
OS/Arch: linux/amd64
Experimental: false

Usually we are able to start new signal and model jobs even when the profile job is running. Every one of them finished.

Sorry for not having yet responded to your colleage. After that troubleshooting we installed a new machine with 5 times more RAM. It didn't block on anything and was really fast in processing. Untill now, when it blocked again. 

Once again thank you for all your time! 

Regards,

Bernardo

PS: In attachment I sent two images. The processors appear to be switching between a state of intense activity and a state of relative calm. It appears the edgems is oscillating between a CPU usage of around 0% (AnalyticsTroubleshootCalm) and of around 400% (AnalyticsTroubleshootNotCalm).

Edit: If you want I can send the metadata of the dataset we are working which blocks on profiles (since my suspicion is that the error has something to do with the java exception thrown as a consequence of having an Integer as Categorical) 

Edit2: We are not using trial version of none 

[pshubham · Accepted solution]

Hi Bernardo,

For this profile job issue, I believe it is related to the salesforce case, that we worked on, where the issue was not reproducible once you re-installed the analytics server. So, can we consider this resolved as of now?

Thanks 

Prakhar Shubham

[BernardoAlb]

Yes, it didn't happen again! 

It is solved, yes!

Thank you, Prakhar 

[pshubham]

Hi Bernardo,

Thanks for the response. I request you to kindly mark my previous answer as the accepted solution. That will save some time for the other community members, who might be experiencing similar kind of issue. We really appreciate your efforts.

Thank you

Best Regards,
Prakhar Shubham