Re: Trident read from Kafka borkers, processes multiple times

2015-03-17 Thread Nikhil Singh
This is interesting.. I am not sure why that would happen    On Friday, March 13, 2015 9:58 AM, Jake Dodd wrote: Looks like you’re using OpaqueTridentKafkaSpout? topology.max.spout.pending, for a Trident spout, refers to the number of in-flight batches. When you set this number too h

Re: Trident read from Kafka borkers, processes multiple times

2015-03-13 Thread Jake Dodd
Looks like you’re using OpaqueTridentKafkaSpout? topology.max.spout.pending, for a Trident spout, refers to the number of in-flight batches. When you set this number too high, OpaqueTridentKafkaSpout sort of freaks out and emits the same tuples in several batches. The batches aren’t failing; th

RE: Trident read from Kafka borkers, processes multiple times

2015-03-13 Thread Qian, Shilei
After I remove the storm configuration "topology.max.spout.pending", the trident workload runs well. But I still get a little confused if I should set this parameter to improve parallelism when processing trident topology. From: Qian, Shilei [mailto:shilei.q...@intel.com] Sent: Tuesday, March 1

Re: Trident read from Kafka borkers, processes multiple times

2015-03-11 Thread Nikhil Singh
This might happen if the batches are failing and replaying over and over again. On Tuesday, March 10, 2015 2:36 AM, "Qian, Shilei" wrote: Hi,   I’m running Storm Trident workload, fetching message from Kafka brokers. Storm version is 0.9.3.   I send just 64 records to Kafka, how