The community will undergo maintenance on October 16th at 10:00 PM PDT and will be unavailable for up to one hour.
Hello everyone,
I am facing the problem, that sometimes when I want to create a model with the thingworx Analytics Builder, the state is stuck in "Running" for days, eventhough similar Models were created much faster. They also dont FAIL.
How can I cancel these Jobs? Delete does not do anything and also restarting the docker Containers of TWx Analytics keep the state in RUNNING.
Any Suggestions?
Best Regards,
Dominik
Solved! Go to Solution.
Can you check out the following directory:
<user-home>/twx-analytics/transfers/training/training/
There will be a master directory with an epoch time. Files should be as: jobStatus.json, jobStatus.touch.json, queuedStatus.json. The directory that matches the timestamp of the job create in Builder is the one you want to delete.
Example:
You will have to remove the directory, then restart the container(s). The Runaway job should then appear as Expired and can be deleted using the Builder delete button.
Alternately, you can try doing it this way:
docker exec -it <training-container-id> /bin/bash cd /tmp/microservices/transfers/training/training grep -R "<job-id>" . # Returns something like "./1527095319573-5/<file>.json". Delete that directory: rm -rf ./1527095319573-5
Then restart the container(s)
Dominik,
Thank you for posting to the Developer Portal.
With the Docker deployment of Analytics, you will have to navigate to the mounted Docker volume and manually delete the parquet references to the run-away job.
On the volume, the Model data is stored in the following directory
/ThingWorxAnalyticsServer/data/data/results/data
You will need to find the file with the same Result ID and Model ID of the model and matching timestamps to that file and delete them manually.
After you do that, you need to restart the Analytics containers, and the run-away jobs will show as "Expired"; and you will be able to remove them using the "Delete" function in Builder, or they will gone already.
Please let me know if that helps.
Regards,
Neel
Hello Neel,
I looked into my running docker containers but could't find your directories. Any Idea where they could be? These are my containers:
This is how they look inside:
Regards,
Dominik
Dominik,
I will have to research this further.
I will be in touch as soon as I obtain the proper directory.
Regards,
Neel
thank you very much!
Can you check out the following directory:
<user-home>/twx-analytics/transfers/training/training/
There will be a master directory with an epoch time. Files should be as: jobStatus.json, jobStatus.touch.json, queuedStatus.json. The directory that matches the timestamp of the job create in Builder is the one you want to delete.
Example:
You will have to remove the directory, then restart the container(s). The Runaway job should then appear as Expired and can be deleted using the Builder delete button.
Alternately, you can try doing it this way:
docker exec -it <training-container-id> /bin/bash cd /tmp/microservices/transfers/training/training grep -R "<job-id>" . # Returns something like "./1527095319573-5/<file>.json". Delete that directory: rm -rf ./1527095319573-5
Then restart the container(s)
Thank you very much, this worked for me!