Thanks. So if you're implementing a Trident Spout, how does that FailedException makes it way onto the spout? I don't see anything in the API of the spout ( https://github.com/apache/incubator-storm/blob/master/storm-core/src/jvm/storm/trident/spout/ITridentSpout.java) that indicates to the spout that the batch needs to replayed. Also does Storm takes of care of cleaning up all the other tuples within this batch that failed as well as any other batches that may have been in the pipeline behind it? Or if we have to implement that, which API would we need to implement?
On Mon, Mar 3, 2014 at 2:52 AM, Nathan Marz <[email protected]> wrote: > Throwing a FailedException is how you programatically fail a batch. > > > On Sun, Mar 2, 2014 at 8:45 AM, David Smith <[email protected]>wrote: > >> Also I don't see a way to a fail batch programmatically like you can do >> with traditional storm? What happens if I throw a Failed Exception from a >> within function, state query or persist? >> >> >> On Sun, Mar 2, 2014 at 9:27 AM, David Smith <[email protected]>wrote: >> >>> I'm trying to implement ITridentSpout but I'm having a hard time figure >>> out where acking for a batch happens. What's the difference between: >>> >>> - ITridentSpout.BatchCoordinator.success >>> - ITridentSpout.Emitter.success >>> >>> What will be called when the whole batch is completed by trident >>> topology? >>> >>> Thanks, >>> David >>> >> >> > > > -- > Twitter: @nathanmarz > http://nathanmarz.com >
