I have a hadoop cluster with 2-master nodes and 3-data nodes. I am running an spark-streaming application. Application is streaming file from s3 and after processing storing result to s3. I'm not able to understand why data node storage getting full without storing data on hdfs. Location: /hadoopdata/hdfs/datanode/current/BP-1127218281-IP-1662958210599/current/finalized/subdir0
Please help how can i handle this situation?