microsoft / Spark-Hive-Tools

This is for issue/feedback tracking on Spark & Hive Tools
Creative Commons Attribution 4.0 International
6 stars 9 forks source link

Pyspark interactive not working when using Spark pool defined in Azure Synapse Workspace #27

Closed erangold closed 2 years ago

erangold commented 3 years ago

I'm trying to use a remote spark cluster, which is defined within a Synapse Workspace environment - I'm able to send batch jobs to Synapse, but when trying to use a pyspark interactive window I get the following error for any command I try to send:

An error was encountered: Invalid status code '500' from https://***.dev.azuresynapse.net/livyApi/versions/2019-11-01-preview/sparkPools/***/sessions/-1 with error payload: {"TraceId":"***","Message":"Something went wrong while processing your request. Please try again later."}

and when I try to access the link within the error I get the following message: {"code":"BearerTokenNotFound","message":"BearerToken not found in request"}

What might be the reason for the 500 error above?

I tried to use the following guide in order to support remote access to Synapse spark pools: Tutorial - Spark & Hive Tools for VSCode (Spark application) - Azure Synapse Analytics | Microsoft Docs

Some technical details:

edit: The guide linked above mentions that this extension doesn't support ms-python versions >= v2020.5.78807, I downgraded to v2020.4.76186 and now I get the following error while trying to initialize a Synapse Pyspark interactive window: Error: Unable to write python.defaultInterpreterPath to Workspace Settings. This setting can be written only into User settings.