-When using other persisters, and simultaneously handling messages which cause changes to the state of the same existing saga instance, NServiceBus ensures that only one message changes the state. Using optimistic concurrency, only one message handler is allowed to succeed. The saga state is updated, and sent messages are dispatched. The other handlers fail, roll back, and their messages enter [recoverability](/nservicebus/recoverability/). When those messages are retried, the process repeats. One or more of those messages may attempt to complete the saga. See below for how that scenario is handled.
0 commit comments