1

The documentation says that any Hadoop API compatible file systems ( like HDFS , S3 ) can be used as checkpoint directory.

My question is that apart from HDFS and S3 what are other practical alternatives for a spark streaming application using Kafka and Cassandra.

Thanks

Soumitra
  • 604
  • 1
  • 8
  • 20

1 Answers1

0

You can use any type of distributed file system like Gluster, GFS, Luster and many more but provided the protocol used by the underlying filesystem should be supported by Spark API's.

Sumit
  • 1,400
  • 7
  • 9