Uploaded image for project: 'Debezium'
  1. Debezium
  2. DBZ-1174

RecordsStreamProducer forgets to convert commitTime from nanoseconds to microseconds

    Details

    • Steps to Reproduce:
      Hide

      1. Start a connector with initial snapshot.
      2. After the snapshot completes, do some manual updates in the monitored tables.
      3. The records from the snapshot have ts_usec in microseconds while the ones from the streaming have ts_usec in nanoseconds.

      Show
      1. Start a connector with initial snapshot. 2. After the snapshot completes, do some manual updates in the monitored tables. 3. The records from the snapshot have ts_usec in microseconds while the ones from the streaming have ts_usec in nanoseconds.

      Description

      RecordsStreamProducer forgets to convert `commitTime` from ReplicationMessage (which is in nanoseconds) to microseconds to match what is being used in the RecordsSnapshotProducer.

      As a result you get the following records with the ts_usec values being inconsistent.

      {"ts_usec":1551879833470000,"txId":105511269,"lsn":5938262159080,"schema":"public","table":
"ims_item","snapshot":true,"last_snapshot_record":false}

{"ts_usec":1551938559907450000,"txId":105513931,"lsn":5941569690428,"schema":"public","table":"ims_item","snapshot":false,"last_snapshot_record":null}

      The issue seems to affect all versions apparently.

        Gliffy Diagrams

          Attachments

            Activity

              People

              • Assignee:
                Unassigned
                Reporter:
                hashhar Ashhar Hasan
              • Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: