Lock manager should expose methods that separate enqueuing a lock request and waiting for it to be satisfied

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Done
    • Priority: Major - P3
    • 2.8.0-rc1
    • Affects Version/s: None
    • Component/s: Concurrency
    • None
    • ALL
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None

      To resolve SERVER-15310 in a race-free manner, it is necessary to enqueue a request for the global shared lock, then kill all outstanding operations, then wait for the lock request to be satisfied. This order is needed to prevent a race between killing outstanding operations and acquiring the global shared lock, during which a new operation might sneak in.

            Assignee:
            Kaloian Manassiev
            Reporter:
            Andy Schwerin
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Created:
              Updated:
              Resolved: