Currently consistency can be tuned by passing config Spark config parameters to the submit job. This could be a configurable option in source/target entries on the config file.
The default consistency for input is LOCAL_ONE, while output is LOCAL_QUORUM. On a cluster that does not run repairs, this means data can be inconsistent among different runs.
Currently consistency can be tuned by passing config Spark config parameters to the submit job. This could be a configurable option in source/target entries on the config file.
Examples of Spark submit parameters:
The default consistency for input is LOCAL_ONE, while output is LOCAL_QUORUM. On a cluster that does not run repairs, this means data can be inconsistent among different runs.