Geek Answers Handbook

How to set the number of map tasks in hadoop 0.20?

I am trying to set the number of map tasks to work in hadoop 0.20 environment.

I am using an old api.

Here are the options I've tried so far:

    conf.set("mapred.tasktracker.map.tasks.maximum", "5");
    conf.set("mapred.map.tasks", "10");
    conf.set("mapred.map.tasksperslot", "5");
    conf.set("mapred.tasktracker.map", "5");
    conf.set("mapred.map.parallel.copies", "5");

With all included, the number of tasks performed in parallel remains equal to 2.

What are the correct parameters for setting the number of parallel running cards to 5?

+5
source share
2 answers

In TaskTracker.java

maxCurrentMapTasks = conf.getInt ("mapred.tasktracker.map.tasks.maximum", 2);

"Hadoop: " . , . .

, . , mapred.tasktracker.map.tasks.maximum , , , , tasktrackers mapred-site.html. , , , mapred.task.tracker.map.tasks.maximum mapred.tasktracker, , tasktracker. , , .

+9

API- Hadoop, .

mapred.tasktracker.map.tasks.maximum => mapreduce.tasktracker.map.tasks.maximum
mapred.map.tasks => mapreduce.job.maps

API API

, .

0

More articles:


All Articles