Hi,
I have several applications using the same hadoop-cluster as data-storage. The different data of each application is stored directly at HDFS or at HDFS by Hive. I want to separate the data of each application “physical” at the cluster like {DataApplication1->DataNode1…5, DataApplication2->DataNode6…10, and so on}. Additional I’m searching for a possibility to grant/denied access to data-nodes like {ServiceUserApplication1->Access to DataNode1…5, ServiceUserApplication2->Access to DataNode6…10, and so on}.
For future the usecase may change to one big data pool for all applications. Then the data will be partitioned by Hive by application (e.g. partitioned by ApplicationID) and the need will be to store partitions at specific data nodes with access to this data nodes (partitoned data) for the ServiceUsers.
To solve the first part I found this Link, but don’t know if it is as dynamic as I need it.
Are there some features to support this usecases?