pvary commented on issue #9089: URL: https://github.com/apache/iceberg/issues/9089#issuecomment-1862676868
The checkpointId is stored in the snapshot summary. It is written only once, and never modified. If you write a new checkpoint, it creates a new snapshot where the summary contains the new checkpointId. The old snapshot with the old checkpointId is still there but when looking for the last checkpointId we go through the history to find the actual one. The logic for finding the current checkpointId is: 1. Check the current snapshot of the table 2. If we find the JobId/OperatorId in the snapshot summary use the checkpointId from here 3. If we do not find it then we look for the previous snapshot in the history and go to step 2 4. If no history left, then we expect that this is the first checkpoint to this table. If someone removes snapshots, then Flink might find missing checkpoints and fail to restart. Finding newer snapshots than the current means that: - Either another Flink job has written to the table - This job was restarted from not the latest checkpoint/savepoint -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
