Uploaded image for project: 'Ansible Automation Platform RFEs'
  1. Ansible Automation Platform RFEs
  2. AAPRFE-1637

Ansible Controller Maintenance mode

XMLWordPrintable

    • False
    • Hide

      None

      Show
      None
    • False

      Comes from: https://issues.redhat.com/browse/ANSTRAT-403

       

      TL;DR Mature users and enterprise organisations need day2 ops style functionality from AAP2 in order to keep operations running smoothly, provide resilience yet but able to cope with things like planned maintenance.

      In AAP2, we are lacking the essential features of a "Maintainance Mode". This Feature is designed to provide the functionality around that.

      Let's say, the user wants to restart the controller or tower services. Due to the nature of the workload, we need to keep the tower alive and do not interrupt any running jobs.

      So putting the controller in maintenance mode, the tower won't schedule jobs on it nor accept any new jobs from users and process the existing running jobs.

      A typical workflow could follow something like this:

      • Admin warns users of pending maintenance (banners, messaging etc)
      • User enables the maintenance mode
        • Stop accepting any jobs from the user
        • Pause any scheduled jobs.
        • Process the existing running jobs
        • Once the existing running jobs are processed, the controller will be in a complete maintenance mode

      Reference - https://www.redhat.com/en/blog/satellite-6-maintenance-mode

      With the right deployment pattern we should aim for a blue-green type deployment model, where one [set] of controllers could enter maintenance mode, the other set take over full automation operations during maintenance.

              bcoursen@redhat.com Brian Coursen
              rhn-support-rafsanch Rafael Minguillón Sánchez
              Votes:
              1 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated: