Windows slaves no longer connect after upgrading to 1.50.3 - ec2-plugin launches many instances requiring manual cleanup

This issue is archived. You can view it, but you can't modify it. Learn more

XMLWordPrintable

    • Type: Bug
    • Resolution: Cannot Reproduce
    • Priority: Major
    • Component/s: ec2-plugin
    • Environment:
      Jenkins 2.235.2 on CentOS 7 (fully patched) running in AWS VPC.
      ec2-plugin: 1.50.3

      We received reports of existing Jenkins masters failing to send build jobs to Windows 2016/2019 slaves upon only upgrading ec2-plugin from 1.50.2.1 to 1.50.3.  We confirmed this by launching a fresh master, applying all updates except ec2-plugin (leaving it at 1.50.2.1).  Then we ran multiple Windows test jobs and saw Windows slaves launched and execute the jobs.  

       

      We then performed only the update of ec2-plugin from 1.50.2.1 to 1.50.3 (nothing else changed) and tried to re-run our test Windows jobs.  This time the Windows slaves did not connect, but reported as "offline" even though I could see they were running in AWS Console, also I could RDP in, also they ran a build just minutes earlier at previous version.

       

      Instead, Jenkins master with the latest ec2-plugin started launching new instances, but never connecting resulting in  10s of instances launched but not used.  Required manual cleanup.

            Assignee:
            FABRIZIO MANFREDI
            Reporter:
            Guy Davis
            Archiver:
            Jenkins Service Account

              Created:
              Updated:
              Resolved:
              Archived: