Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-35057

Notify oplog waiter after advancing the lastApplied timestamp

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 4.0.0-rc2, 4.1.1
    • Component/s: Storage
    • Labels:
      None
    • Backwards Compatibility:
      Fully Compatible
    • Operating System:
      ALL
    • Backport Requested:
      v4.0
    • Sprint:
      Storage NYC 2018-06-04, Storage NYC 2018-06-18

      Description

      We create a new test suite called concurrency_replication_causal_consistency inĀ SERVER-32883, which runs all the fsm workloads with causal consistency on a 5-node linear chained replica set.
      Tests such as auth_create_role.js timed out due to very slow dropRole operations which are done with majority write concerns. We later figured out the slowness was caused by not notifying the oplog waiter immediately after advancing the lastApplied timestamp. So another secondary which syncs from the secondary would wait long time(~20s) for the notification that lastApplied has been advanced.

        Attachments

          Issue Links

            Activity

              People

              Assignee:
              xiangyu.yao Xiangyu Yao (Inactive)
              Reporter:
              xiangyu.yao Xiangyu Yao (Inactive)
              Participants:
              Votes:
              0 Vote for this issue
              Watchers:
              8 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved: