apache / apache/beam

Fix context classloader in Spark portable runner

Open
#20,192 0 comments 0 reactions 0 assignees View on GitHub
improvement P3 portability runners spark
Dominant language
Java
Stars
8.7k
Forks
4.7k
Avg merge
1d 20h
Merged PRs (30d)
196

Description

Spark is setting the context class loader to support dynamic class loading, leading to unpredictable behavior with duplicate jars being found on the class path. We need to see if there is a way to disable this behavior so we can use the context class loader deterministically.

Imported from Jira [BEAM-10109](https://issues.apache.org/jira/browse/BEAM-10109). Original Jira may contain additional context.
Reported by: ibzib.

Contributor guide

Open the contributing guide

Research direction

Start by reading the Spark portable runner implementation and the linked BEAM-10109 Jira issue for the missing context. Determine how Spark sets the context class loader and whether that behavior can be disabled; done means duplicate jars no longer cause nondeterministic class loading.

Written by the indexing model from the issue text.

Assessment

Tech stack
java, spark
Domain
distributed-systems
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.