Question

I am install hadoop on signal node and run program but delay time if have dataset size 3 3 GB take time about 2 hours

Answer 1

Hey @Faris!

Try increasing the number of reducers. When you increase the number of reducers, the distribution is better. I have seen in a few cases that only 1 reducer is used and this causes delay. You can set the number of reducers while running the MapReduce program. Refer to the below command:

$ hadoop <the options that you usually pass> -D mapred.reduce.tasks=<number of reducers>

Comments

Mean increasing number slaves?

---

No, I mean the reducer. You are running a mapreduce program. And the job will be divided to mappers and reducers. So, I think increasing the number of reducers might help.

---

How i can increasing number reduce ;please help me

---

1. Export your MapReduce Java project to JAR file. (Refer to this: https://www.tutorialspoint.com/eclipse/eclipse_create_jar_files.htm)
2. Open terminal and go to the directory where you have exported the JAR file. 
3. Run the MapReduce using the following command syntax:

$ hadoop jar <jar file name> <Java class name with packagename(Ex:com.hadoop.WordCount)> <hdfs input path> <hdfs output path>