Uploaded image for project: 'DC/OS'
  1. DC/OS
  2. DCOS_OSS-448

How to rebuild minuteman and navstar config file on a failed node ?

    Details

      Description

      Hi, from time to time, I've been through the following problem:
      navstar and / or minuteman crashes and fail to restart because of vm.args env parameter missing. This is bad because it make the whole service discovery to fail in my cluster.

      I've just found out that my nodes were filling up with unused docker images, leading to no disk space available and corruption of two files: sys.config and vm.args in either or both directories:

      /opt/mesosphere/packages/minuteman--<your_package_number_here>/minuteman/releases/0.0.1/
      /opt/mesosphere/packages/navstar--<your_package_number_here>/navastar/releases/0.1.0/
      

      Result of corruption: empty files.
      Copying the sys.config and vm.args from other nodes and editing the vm.args (it is hardcoding the host ip address) was enough to have the services restarting, but I was wondering how should I do to regenerate thoses two files using the installation scripts ?

      Cheers

      David

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                dgoel Deepak Goel
                Reporter:
                davvdgign David Vandergucht (Inactive)
                Team:
                Networking Team
                Watchers:
                David Vandergucht (Inactive), Deepak Goel, olek, Sergey Urbanovich
              • Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: