[ https://issues.apache.org/jira/browse/BEAM-8151?focusedWorklogId=306738&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-306738 ]
ASF GitHub Bot logged work on BEAM-8151: ---------------------------------------- Author: ASF GitHub Bot Created on: 04/Sep/19 21:55 Start Date: 04/Sep/19 21:55 Worklog Time Spent: 10m Work Description: aaltay commented on pull request #9477: [BEAM-8151, BEAM-7848] Up the max number of threads inside the SDK harness to a default of 10k URL: https://github.com/apache/beam/pull/9477#discussion_r320992748 ########## File path: sdks/python/apache_beam/runners/worker/sdk_worker_main.py ########## @@ -202,7 +202,7 @@ def _get_worker_count(pipeline_options): re.match(r'worker_threads=(?P<worker_threads>.*)', experiment).group('worker_threads')) - return 12 + return 10000 Review comment: Do we know why 10K is a good number? Would not this practically accept all work? How did we come up with 12 in the first place. (@angoenka may have information) ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 306738) Time Spent: 1h 20m (was: 1h 10m) > Allow the Python SDK to use many many threads > --------------------------------------------- > > Key: BEAM-8151 > URL: https://issues.apache.org/jira/browse/BEAM-8151 > Project: Beam > Issue Type: Improvement > Components: sdk-py-core, sdk-py-harness > Reporter: Luke Cwik > Assignee: Luke Cwik > Priority: Major > Time Spent: 1h 20m > Remaining Estimate: 0h > > We need to use a thread pool which shrinks the number of active threads when > they are not being used. > > This is to prevent any stuckness issues related to a runner scheduling more > work items then there are "work" threads inside the SDK harness. > > By default the control plane should have all "requests" being processed in > parallel and the runner is responsible for not overloading the SDK with too > much work. -- This message was sent by Atlassian Jira (v8.3.2#803003)