To get started, our requirement is quite simple when tweets occur, then only release (only at regular intervals) on HDFS Need to keep
JAVSTREAMING CONTACTEC promises 'Checkpoint API', but after a further review it is a different purpose. (Apart from this, I do not keep '/ checkpoint / temporary, error: such as a file or directory (2)' error, but do not worry about it yet.) Question: JavadiTrim does not have the 'sehadopfile' method - which makes little sense. I think saving Hodop from a streaming job is not a good idea.
What is the recommended approach? Should I write the incoming 'tweet' in the kafka queue and then use the tool like 'camus' () to push HDFS?
This awesome blog entry came in to confirm my views. Using technologies like Kafka, Hurricane, Camus, created a 'foreign currency trading system'. This use is like mine, so I'm going to use this design; equipment. Thank you.
Comments
Post a Comment