Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-43628

Jenkins slaves were started and stopped during master restart

    Details

    • Type: Bug
    • Status: Fixed but Unreleased (View Workflow)
    • Priority: Minor
    • Resolution: Fixed
    • Component/s: ec2-plugin
    • Labels:
    • Environment:
      Jenkins Versions: 2.7.4 & 2.60.3
      Amazon EC2 plugins: 1.35 & 1.36
      Idle termination time: -5
    • Similar Issues:
    • Released As:
      1.41

      Description

      After Jenkins master was restarted, the software was looping trying to connect to the slaves through SSH.  When I checked AWS, I saw the slaves were in a stopped state.  This continued on until the LaunchTimeout was reached.  It looks like Jenkins both started and stopped the slaves during restart.  Below is the output from the log:

      11:45:35 PM - Jenkins was restarted through the app

      Jenkins slaves started on startup
      11:45:36 PM hudson.plugins.ec2.EC2RetentionStrategy start
      INFO: Start requested for JENKINS_SLAVE_NAME (AWS_INSTANCE_ID)
      11:45:38 PM hudson.plugins.ec2.EC2Cloud log
      INFO: Launching instance: AWS_INSTANCE_ID

      Jenkins slaves stopped?
      INFO: Idle timeout of JENKINS_SLAVE_NAME (AWS_INSTANCE_ID) after 0 idle minutes, with 0 minutes remaining in billing period
      11:45:39 PM hudson.plugins.ec2.EC2AbstractSlave idleTimeout
      INFO: EC2 instance idle time expired: AWS_INSTANCE_ID
      11:45:39 PM hudson.plugins.ec2.EC2AbstractSlave stop

      After slaves were stopped, Jenkins reports as up
      11:45:47 PM hudson.WebAppMain$3 run
      INFO: Jenkins is fully up and running

      Jenkins unable to connect to stopped slaves
      11:45:53 PM hudson.plugins.ec2.EC2Cloud log
      INFO: Connecting to IP_ADDRESS on port 22, with timeout 10000.

      11:46:09 PM hudson.plugins.ec2.EC2Cloud log
      INFO: Connecting to IP_ADDRESS on port 22, with timeout 10000.

      11:46:24 PM hudson.plugins.ec2.EC2Cloud log
      INFO: Connecting to IP_ADDRESS on port 22, with timeout 10000.
      ...
      3:19:55 AM hudson.plugins.ec2.EC2Cloud log
      INFO: Connecting to IP_ADDRESS on port 22, with timeout 10000.

        Attachments

          Issue Links

            Activity

            Hide
            thoulen FABRIZIO MANFREDI added a comment -

            In the 1.39 has been solved the problem of stopped node, in 1.41 there are several improvements on stopped nodes

            Show
            thoulen FABRIZIO MANFREDI added a comment - In the 1.39 has been solved the problem of stopped node, in 1.41 there are several improvements on stopped nodes

              People

              • Assignee:
                thoulen FABRIZIO MANFREDI
                Reporter:
                mkozell Mike Kozell
              • Votes:
                2 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: