Open
Description
I dunno if there is anything obvious that springs to mind here @thammegowda or @karanjeets from back in the day.
When I run Sparkler as a spark submit job on a databricks cluster, even with the partitioned RDD's all the tasks are executed on the master node
You can see here, the workload is executed on the master node, but not pushed to the other 3. Which seems entirely reproducable.
Metadata
Metadata
Assignees
Labels
No labels