Getting MemoryError when executing task hdfs dfs -format; how to fix / troubleshoot?

We have a task in our Ansible based deploy which performs formatting of HDFS using the hdfs dfs -format command.

This command causes a MemoryError. We can see that there is a Python process on the target box which gobbles up ~32GB of memory and then the failure occurs.

How can we troubleshoot this further?

Also, any ideas/recommendations on how to manage Ansible’s memory usage? I’m peeking into ansible_memtotal_mb… any config settings to look at or adjust?

Thanks,

  • Dmitry