Set catchUpTimeoutMillis if catchup takeover is disabled

XMLWordPrintableJSON

    • Type: Improvement
    • Resolution: Unresolved
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Replication
    • None
    • Replication
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None

      By default `catchUpTimeoutMillis` is set to -1 (infinity). If user disabled catchup takeover, without setting `catchUpTimeoutMillis`, and for whatever reason the primary is stuck in catchup/drain mode, then the system can freeze. Although users can manually issue `replSetAbortPrimaryCatchUp` command to abort catchup, this is still not good user experience because it involves user intervention and long unavailability window. We should either warn users when they disable catchup takeover without setting `catchUpTimeoutMillis` or alternatively fail the configuration when that happens.

            Assignee:
            [DO NOT USE] Backlog - Replication Team
            Reporter:
            Wenbin Zhu
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated: