Guaranteed processing
Storm provides an API that allows you to guarantee that a tuple emitted by a spout is fully processed. So far in our example, we've not worried about failures. We've seen that a spout stream can be split and can generate any number of streams in a topology, depending on the behavior of downstream bolts. What happens in the event of a failure? As an example, consider a bolt that persists information to tuple data based on a database. How do we handle situations where the database update fails?
Reliability in spouts
In Storm, guaranteed message processing begins with the spout. A spout that supports guaranteed processing needs a way to keep track of tuples it has emitted and be prepared to re-emit a tuple if downstream processing ...
Get Storm Blueprints: Patterns for Distributed Real-time Computation now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
