Uploaded image for project: 'Apache YuniKorn'
  1. Apache YuniKorn
  2. YUNIKORN-553

Gang scheduling enhancements

Attach filesAttach ScreenshotAdd voteVotersWatch issueWatchersCreate sub-taskLinkCloneUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

    Details

      Description

      This is an umbrella jira that tracks all enhancements for the gang scheduling feature YUNIKORN-2.

        Attachments

        Issue Links

        1.
        Add UT coverage for YUNIKORN-460 Sub-task Resolved Kinga Marton Actions
        2.
        Rename scheduling parameter placeholderTimeout to placeholderTimeoutInSeconds Sub-task Resolved Kinga Marton Actions
        3.
        Add some unit tests to cover placeholder cleanup Sub-task Open Kinga Marton Actions
        4.
        Placeholder pods are not cleaned up even when the job is deleted Sub-task Resolved Kinga Marton Actions
        5.
        Publish events to app's pods if the app is failed to be scheduled Sub-task Resolved Unassigned Actions
        6.
        Remove the WARN log when placeholderTimeout is not defined in taskGroups Sub-task Resolved Ting Yao,Huang Actions
        7.
        Yunikorn recovery deletes existing placeholders Sub-task Open Kinga Marton Actions
        8.
        Add a label to the placeholder pods the selector Sub-task Resolved Ting Yao,Huang Actions
        9.
        Make sure all timer go routines are stopped after removing an app Sub-task Open Manikandan R Actions
        10.
        Check if the placeholders and real resources are the same Sub-task Open Manikandan R Actions
        11.
        Gang scheduling Design doc Sub-task Resolved Wilfred Spiegelenburg Actions
        12.
        Update state machine doc Sub-task Resolved Kinga Marton Actions
        13.
        Consider a fallback mechanism to schedule the app incase of gang failure instead of rejecting the app. Sub-task Open Manikandan R Actions
        14.
        Create a Failing app state in shim side Sub-task Open Manikandan R Actions
        15.
        Placeholder pods are not cleaned up timely when the Spark driver fails Sub-task Open Kinga Marton Actions
        16.
        Allocated resources on a node could become negative Sub-task Resolved Chaoran Yu Actions
        17.
        Optimize the UT for ListApplications Sub-task Open Weiwei Yang Actions
        18.
        Expose pod level events when an app is failed in scheduling Sub-task Resolved Ting Yao,Huang Actions
        19.
        Gang scheduling User Guide Sub-task Resolved Weiwei Yang Actions
        20.
        Enhance placeholder cleanup on timeout Sub-task Resolved Wilfred Spiegelenburg Actions
        21.
        Publish a pod event to indicate the task is being gang scheduling Sub-task Resolved Ting Yao,Huang Actions
        22.
        enhanced StateAware gang scheduling Sub-task Resolved Wilfred Spiegelenburg Actions
        23.
        Add events for placeholder timeout to pod Sub-task Open Manikandan R Actions
        24.
        Make placeholder image configurable Sub-task Open Amit Sharma Actions

          Activity

            People

            • Assignee:
              Unassigned
              Reporter:
              wwei Weiwei Yang

              Dates

              • Created:
                Updated:

                Issue deployment