Details
-
Bug
-
Status: Resolved
-
Blocker
-
Resolution: Fixed
-
0.9.0
-
None
Description
It would be useful to have a script, roughly similar to the "hadoop jar" command that is used for running compiled binaries against Spark.
The script would do two things:
- Set up the Spark classpath on the client side, so that users don't need to know where Spark jars are installed or bundle all of Spark inside their app jar.
- Provide a layer over the different modes that apps can be run, so that the same spark-jar invocation could run the driver inside a YARN application master or in the client process, depending on the cluster setup.
Attachments
Issue Links
- is related to
-
SPARK-1318 Alignment of the Spark Shell with Spark Submit.
- Resolved