This topic provides the steps to enable execute the mapping in the streaming mode. Streaming needs checkpointing information for a fault-tolerant storage system to recover from failures.
- Click the Topology tab.
- In the Physical Architecture tree, under Technologies, right-click Spark Python and then click New Data Server.
- In the Definition tab, specify the details of the Spark data server.
- In the Properties tab, specify the properties for the Spark data server.
- Create a mapping using the data store.
- Enable the Streaming flag in the Physical design of a mapping.
- Enable the
spark.checkpointing
to set the Checkpointing directory.
Every mapping will have its unique checkpointing directory.
- Execute the mapping and set the context for physical design.
Note:
In the
User Interface Designer by default, the
Last executed physical design in the mapping execution dialog is pre-selected.