Currently, Oozie can look at data dependencies in hcat, hdfs, but do not have support for kafka topics. You can have a java action which can keep on checking for topic in kafka (probably using some REST APIs available in kafka. I don't know much about kafka) and once the topic is available go to the spark action.
On Fri, Dec 2, 2016 at 2:38 PM, prateek arora <[email protected]> wrote: > Hi team , > > is it possible to launch a spark job using oozie when data is available in > kafka topic ? > > Regards > Prateek >
