Details
Major Description
I did apply HUE-4761 to Hue 3.11.0, and I'm experiencing the following behaviour:
adding jars to the session works fine the first time I'm recreating the session, but the next times the modification are not taking into account, the session recreation are always using the jars from the first time.
Dumping the Livy REST call, I get:
First session creation, no jars set => correct.
method=POST url=http://localhost:8998/sessions params=None data={"files": [], "pyFiles": [], "kind": "spark", "proxyUser": "bperroud", "driverMemory": "1G", "queue": "default", "archives": [], "executorCores": 1, "driverCores": 1, "jars": [], "executorMemory": "1G"} headers={'Content-Type': 'application/json'}
Adding one jar, recreating the session with one jar => correct
method=POST url=http://localhost:8998/sessions params=None data={"files": [], "pyFiles": [], "kind": "spark", "proxyUser": "bperroud", "driverMemory": "1G", "queue": "default", "archives": [], "executorCores": 1, "driverCores": 1, "jars": ["/DataScienceLab/libraries/spark-csv_2.10-1.5.0.jar"], "executorMemory": "1G"} headers={'Content-Type': 'application/json'}
Adding a second jar, recreating the session. The list of jars is not update => incorrect.
method=POST url=http://localhost:8998/sessions params=None data={"files": [], "pyFiles": [], "kind": "spark", "proxyUser": "bperroud", "driverMemory": "1G", "queue": "default", "archives": [], "executorCores": 1, "driverCores": 1, "jars": ["/DataScienceLab/libraries/spark-csv_2.10-1.5.0.jar"], "executorMemory": "1G"} headers={'Content-Type': 'application/json'}
As a workaround, saving the notebook and reloading the page (notebook) produce the correct result.
Thanks