Uploaded image for project: 'Apache Drill'
  1. Apache Drill
  2. DRILL-4186 queue enhancements [umbrella]
  3. DRILL-4202

Handle coordinator/foreman failures to prevent data loss

    XMLWordPrintableJSON

Details

    • Sub-task
    • Status: Open
    • Minor
    • Resolution: Unresolved
    • None
    • None
    • None
    • None

    Description

      Foreman relies on ephemeral nodes to get rid of zombie profiles. However, foreman failures still cause loss of profile data. This happens at any non-terminal state where profile is not yet persisted. The initial proposal is to rely on watchers to detect state changes and react upon. We can use random back-off or a similar scheme to avoid hammering Zookeeper.

      Attachments

        Activity

          People

            Unassigned Unassigned
            hgunes Hanifi Gunes
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

            Dates

              Created:
              Updated: