[FLINK-20933] Config Python Operator Use Managed Memory In Python DataStream

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.12.0, 1.13.0
• Fix Version/s: 1.12.2, 1.13.0
• Component/s: API / Python
• Labels: pull-request-available
• Assignee: Huang Xingbo
• Reporter: Huang Xingbo
• Created: 12/Jan/21 04:40
• Updated: 28/May/21 07:10
• Resolved: 18/Jan/21 05:09

Description

Now the way to set `Python DataStream Operator` to use managed memory is to set a hook in the `execute` method of `Python StreamExecutionEnvironment` to traverse the `StreamGraph` and set the `Python Operator` to use managed memory.
But when the user’s job uses `from_data_stream` to convert the `DataStream` to a `Table`, the `TableEnvironment.execute` method is used at the end rather than `StreamExecutionEnvironment.execute`, so the `Python DataStream` related operators will not have `Managed Memory` set.

Issue Links

• links to: GitHub Pull Request #14621
• links to: GitHub Pull Request #14676