0xdeafbeef opened a new pull request, #2780: URL: https://github.com/apache/datafusion-ballista/pull/2780
# Which issue does this PR close? No issue filed; the bug is described below. # Rationale for this change The shuffle writers do their work in `tokio::spawn` tasks that nothing joins or aborts, so cancelling a task leaves them running until the input ends. A cancelled task keeps burning CPU. # What changes are included in this PR? The coordinator and the tasks of `drive_shuffle_writer_stage` are now DataFusion's `SpawnedTask`, which aborts on drop. The `execute(N)` streams own the coordinator through an `Arc`; the writer state keeps only a `Weak`, because the executor keeps the plan alive after a cancel. Along the way, the free `take_handoff` became a method of `WriterState`, and `SortShuffleWriterExec` uses the shared `WriterState` instead of its own identical copy. # Are there any user-facing changes? A cancelled task no longer burns CPU. No API changes. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
