Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-72194

oplog_rollover.js should wait for the durable timestamp to advance

    • Type: Icon: Bug Bug
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • 6.3.0-rc0, 6.0.5, 4.4.20, 5.0.16
    • Affects Version/s: None
    • Component/s: None
    • Labels:
    • Fully Compatible
    • ALL
    • v6.0, v5.0, v4.4
    • Repl 2022-12-26, Repl 2023-01-09, Repl 2023-01-23, Repl 2023-02-06
    • 14

      The durable timestamp on the primary is also a pinned timestamp preventing oplog truncation, so we need to wait for this advance before expecting truncation to happen. On the regular wiredTiger storage engine this isn't an issue because the recovery timestamp we are checking for is always durable, but on the ephemeral engine (where "durable" is fake), the "recovery timestamp" is actually just the majority point and may not have been marked "durable".

      It should be sufficient to wait for durableOpTime to advance on the primary.

            matthew.russotto@mongodb.com Matthew Russotto
            matthew.russotto@mongodb.com Matthew Russotto
            0 Vote for this issue
            4 Start watching this issue