Setting the Map/Reduce Memory on the Client
Symptom
How do I set Map and Reduce memory on the client?
Procedure
Before SQL statement execution, run the set command to set parameters of clients related to Map/Reduce.
The following parameters are related to Map and Reduce memory:
set mapreduce.map.memory.mb=4096; //Memory required by each Map taskset mapreduce.map.java.opts=-Xmx3276M; //Maximum memory used by the JVM of each Map taskset mapreduce.reduce.memory.mb=4096; //Memory required by each Reduce taskset mapreduce.reduce.java.opts=-Xmx3276M; //Maximum memory used by the JVM of each Reduce taskset mapred.child.java.opts=-Xms1024M -Xmx3584M; // This parameter is a global parameter, which is used to set Map and Reduce in a unified manner.
Note
Parameter settings take effect for the current session only.
Parent topic: Using Hive
- Symptom
- Procedure