Flink TaskManager Memory

[Timothy Victor]

For Streaming Jobs that use RocksDB my understanding is that state is allocated off-year via RocksDB.

If this is true then does it still make sense to leave 70% (default taskmanager.memory.fraction) of the heap for Flink Manged memory given that it is likely not being used for state?    Or am I missing something, and it is in fact still used?

Thanks

Tim

[vino yang]

Hi Tim,

Reference a blog comes from Ververica:

"When you choose RocksDB as your state backend, your state lives as a serialized byte-string in either the off-heap memory or the local disk."

It also contains many tune config options you can consider.[1]

Best,

Vino

[1]: https://www.ververica.com/blog/manage-rocksdb-memory-size-apache-flink

Timothy Victor <[hidden email]> 于2019年12月27日周五 上午7:11写道：

For Streaming Jobs that use RocksDB my understanding is that state is allocated off-year via RocksDB.

If this is true then does it still make sense to leave 70% (default taskmanager.memory.fraction) of the heap for Flink Manged memory given that it is likely not being used for state?    Or am I missing something, and it is in fact still used?

Thanks

Tim