Details

    • Type: Improvement
    • Status: Closed (View Workflow)
    • Priority: Major
    • Resolution: Not A Defect
    • Component/s: ec2-plugin
    • Labels:
      None
    • Environment:
      Ubuntu Linux, Jenkins 2.176.1, EC2 plugin 1.43
    • Similar Issues:

      Description

      Great plug-in, thank you so much for all the great work!

      The time interval from when a job lands in the build queue until a new slave node is started to handle the job is rather long.  It would be useful if we could tune how frequently the build queue is checked and how quickly new slaves are started up so we can reduce the wait time.

        Attachments

          Activity

          Hide
          raihaan Raihaan Shouhell added a comment -

          There is an option for nodelayprovisioning which starts slaves as soon as required without waiting. The standard strategies have logic around when to start slaves.

          Show
          raihaan Raihaan Shouhell added a comment - There is an option for nodelayprovisioning which starts slaves as soon as required without waiting. The standard strategies have logic around when to start slaves.
          Hide
          thoulen FABRIZIO MANFREDI added a comment -

          Don't forget that to keep the OS small enough to be fast to boot, the generic algorithm used to raise new node is already quite fast. How much is your startup time ?  

          Show
          thoulen FABRIZIO MANFREDI added a comment - Don't forget that to keep the OS small enough to be fast to boot, the generic algorithm used to raise new node is already quite fast. How much is your startup time ?  
          Hide
          thoulen FABRIZIO MANFREDI added a comment -

          With the structure that we have we are not able to reduce the startup time, already we start parallel node base on the queue size. More operation will result in a deadlock of the queue. 

          Show
          thoulen FABRIZIO MANFREDI added a comment - With the structure that we have we are not able to reduce the startup time, already we start parallel node base on the queue size. More operation will result in a deadlock of the queue. 

            People

            • Assignee:
              thoulen FABRIZIO MANFREDI
              Reporter:
              daschneider99 D.Schneider
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: