Since we will turn a retryable write into a transaction when an update to the shard key is sent as a retryable write, we need to migrate some info during session migration about this write. That way, if the chunk that owned the original document moves, an attempted retry will fail. In order to do this, we will need to add some extra field to config.transactions to indicate that the retryable write turned into a transaction.
- related to
-
SERVER-51750 config.transactions cloner for resharding must not ignore records for multi-document transaction
- Closed
-
SERVER-38844 SessionMigrationSource should explicitly ignore config.transactions entry meant for transactions
- Closed