[
https://issues.apache.org/jira/browse/BEAM-8816?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anonymous updated BEAM-8816:
----------------------------
Status: Triage Needed (was: Resolved)
> Load balance bundle processing w/ multiple SDK workers
> ------------------------------------------------------
>
> Key: BEAM-8816
> URL: https://issues.apache.org/jira/browse/BEAM-8816
> Project: Beam
> Issue Type: Improvement
> Components: runner-core, runner-flink
> Affects Versions: 2.17.0
> Reporter: Thomas Weise
> Assignee: Thomas Weise
> Priority: P2
> Labels: portability
> Fix For: 2.19.0
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> We found skewed utilization of SDK workers causing excessive latency with
> Streaming/Python/Flink. (Remember that with Python, we need to execute
> multiple worker processes on a machine instead of relying on threads in a
> single worker, which requires the runner to make a decision to which worker
> to give a bundle for processing.)
> The Flink runner has knobs to influence the number of records per bundle and
> the maximum duration for a bundle. But since the runner does not understand
> the cost of individual records, it is possible for the duration of bundles to
> fluctuate significantly due to skew in processing time of individual records.
> And unless the bundle size is 1, multiple expensive records could be
> allocated to a single bundle before the cutoff time is reached. We notice
> this with a pipeline that executes models, but there are other use cases
> where the cost of individual records can vary significantly.
> Additionally, the Flink runner establishes the association between the
> subtask managing an executable stage and the SDK worker during
> initialization, lasting for the duration of the job. In other words, bundles
> for the same executable stage will always be sent to the same SDK worker.
> When the execution time skew is tied to specific keys (stateful processing),
> it further aggravates the issue.
> [https://lists.apache.org/thread.html/59c02d8b8ea849c158deb39ad9d83af4d8fcb56570501c7fe8f79bb2@%3Cdev.beam.apache.org%3E]
> Long term this problem can be addressed with SDF. Till then, an (optional)
> runner controlled balancing mechanism has shown to improve the performance in
> internal testing.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
