THIS WAS TESTED ON SANDBOX HDP 2.3
- start kafka in ambari a. there is sometimes a glitch will it not show as started b. go to the spark configuration settings in ambari
- under advanced spark-log4j-properties cahnge “log4j.rootCategory” to be equal to “ERROR” a. (Re)Start spark
- start deployment script "SH deployment_script.sh"
- in sandbox add service, select nifi and next
- Go with all default options, if errors are on customize services a. remove oozie.authentication.kerberos.name.rules with the red button to the right b. set the ranger db root password to anything
- In virtual box go to settings>network>port forwarding and add port 9090 for nifi
- separately download “streaming_demo.xml” from the repo and upload as a template in nifi
- start nifi allprocessors, a few aren't configured to work but event gen through hdfs should work
- start spark streaming with “sh kafka_test.sh”