Hello everyone, my topology contains only one spout. It fetches 16000 records from mySQL, stores them in a blocking queue created by this spout, then emits them one by one. As soon as the queue becomes empty, the spout will fetch 16000 records again from mySQL, and start to do the same iteration above. The eimtted tuples will be processed by one bolt. For the spout we have set just one executor, and for bolt we set 15 executors and 30 tasks. The processed data will be transfered to next bolt. It stores the result in HBase.
At the beginning all tuples were processed very quickly, in 40-50 mins. But after 10-12 iterations, the processing becomes quite slowly. In last 3 hours were only 6000 processed by bolts. What could be the reason? any ideas would be appreciated. Jun
