Скачать или смотреть Using AWS Elastic MapReduce EMR as a Generic Hadoop Cluster to make a Word Counter - Ing Alex Tapia

// USING MAPREDUCE ON THE AMAZON CLOUD

BUILT THE CLUSTER IN THE CLOUD
1 make a key pair
2 Create S3 bucket
You need to have a disk drive in the cloud where you can put
a. your input data,
b. you can collect your output data
c. your code
3. Create a MapReduce Cluster EMR (Elastic Map Reduce), a Managed Hadoop Framework
This is being completed well, so we are going to switch back to our S3 interface and look in our bucket.
In the S3, We refresh and see that a new output folder it's been created. It's one file for each reduce. each one of this is a sorted list of terms that were output by a reducer class
Download it, open it up, we can see that we have lots o lots of terms and the number of times that they occured
IMPORTANT
The last thing to do is to shut down the cluster cleanlly to make sure you are not pang for the cluster when you are not using it
1. To do that, you wanna go to your MapReduce S3 bucket, and you wanna make sure that you are not storing anything in it.
If you actually want the output, you need to download it, then select all the files that are present, the JAR and the input file, then delet ,them, this way you are not charge
2. Make sure that your cluster is shut down from EC2, So here you see that are running 3 instances, I'm gonna go and select that and you can see that 3 computers are still running.
I want to select those 3 instances and check and change my Action, en Instance State, select Terminate
So As long as I don't have any EC2 computers running and as long as don't have any files in my S3 bucket, I won't be charge anymore
Source: Prof. Patterson shows how to get a Java program running in the Hadoop Map/Reduce framework used by Amazon's Web Services platform.
   • Creating a Java Program for Map/Reduce (Pa...