Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-69700

Replication rollback fails resharding_coordinator_recovers_abort_decision.js

    • Fully Compatible
    • ALL
    • v6.1, v6.0, v5.0
    • Sharding 2022-09-19, Sharding 2022-10-03
    • 35
    • 1

      In BF-26341, the following can be observed in logs:


      [js_test:resharding_coordinator_recovers_abort_decision] c20296| 2022-09-11T05:28:13.563+00:00 I  REPL     21098   [BackgroundSync] "Starting rollback due to fetcher error","attr":{"error":"OplogStartMissing: the sync source's oplog and our oplog have diverged, going into rollback. our last optime fetched: { ts: Timestamp(1662874093, 2), t: 1 }. optime of first document in batch: { ts: Timestamp(1662874093, 4), t: 2 }. sync source's first optime: { ts: Timestamp(1662874088, 18), t: 1 }","lastCommittedOpTime":{"ts":{"$timestamp":{"t":1662874092,"i":42}},"t":1}}
      [js_test:resharding_coordinator_recovers_abort_decision] c20297| 2022-09-11T05:28:13.980+00:00 I  REPL     21105   [BackgroundSync] "Rollback successful"

      This causes the test client disconnection and the command to turn off the failpoint fails. A fix here would be to wait for secondaries to be ready before doing so.


            abdul.qadeer@mongodb.com Abdul Qadeer
            abdul.qadeer@mongodb.com Abdul Qadeer
            0 Vote for this issue
            3 Start watching this issue