Allow single KT to run and not automatically destroy resourcesSee original GitHub issue
An iteration of the harness runner for a single KT destroys its resources at the end of its run. This totally makes sense that it would work this way to validate the KT.
However, this impacts the workflow for the user who wants to run the KT and leave it running to interrogate the system. This would be a blocker for someone who wants to leave it running and do other steps, query it, etc.
The cleanup happens in the last step of all harnesses, e.g., https://github.com/confluentinc/kafka-tutorials/blob/master/_data/harnesses/filtering/ksql.yml#L106
Harness-runner can take a 3rd parameter, the stages to run. By default it’s dev, test, prod. https://github.com/confluentinc/kafka-tutorials/blob/master/_includes/tutorials/filtering/ksql/code/Makefile#L12
If we switched line 22 to below, it could work:
harness-runner ../../../../../_data/harnesses/filtering/ksql.yml $(TEMP_DIR) dev,test
This is where the envs get selected. https://github.com/confluentinc/kafka-tutorials/blob/master/harness_runner/harness_runner.py#L134
- Created 3 years ago
- Comments:6 (6 by maintainers)
Top GitHub Comments
I vote for Option 1 above. If a second Makefile target is desired, it could be added but it just delegates to the
tutorialtarget passing in the
SEQUENCEparameter as needed. That would eliminate the duplication across two targets
Addressed via https://github.com/confluentinc/kafka-tutorials/pull/454