Skip to content

Duplicate channels can crash range shuffle routing #8245

Description

@carloea2

Feature Summary

Python range shuffle divides its key range by the channel count but routes through a deduplicated receiver list. The planner creates every sender-receiver pair, so multiple senders can make a valid key produce an out-of-bounds receiver index.

Configuration Before After
Two senders, two receivers, range -10 to 9, key 9 Four channels give bucket 3 and IndexError Two receivers give bucket 1, matching the JVM

Reproduced on upstream main 289b70945bdd8ec5b4a10016a1ecebe898d931ef by executing LinkConfig.toPartitioning, serializing its result, and passing it to Python OutputManager. Both single-sender controls pass; all four cases with two or three senders and one or two receivers fail before the change.

Proposed Solution or Design

Divide the key range by the unique receiver count already used for routing. Add an OutputManager regression with the planner's complete sender-receiver topology, including boundary and out-of-range values.

Affected Area

Workflow Engine (Amber)

Activity

  1. added theissue type on Aug 31, 2026
  2. changed the issue type fromtoon Aug 31, 2026
  3. carloea2 commented on Sep 5, 2026

    @carloea2
    ContributorAuthor

    /take

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions