Fix updating optimes from remote nodes

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Won't Fix
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Replication
    • None
    • ALL
    • 0
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Currently, it is possible for a node to erroneously update its own lastAppliedOptime when all of these conditions are true:

      • the node is PRIMARY
      • the node was SECONDARY in the past and has chained through another secondary
      • the node does further writes before receiving a handshake from the chaining secondary

      In such a situation, we rewind our own optime in the slavemap, but do not update the legacy replset's lastappliedoptime, which causes an fassert failure.

            Assignee:
            Eric Milkie
            Reporter:
            Eric Milkie
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: