hadoop-ec2 in Hadoop package: how to specify a region?

Go To StackoverFlow.com

1

I want to launch Hadoop cluster on EC2 in some Asian AWS region but it is being set up in East Coast one (Virginia) by default and I can't find any relevant parameter or environment variable for that from bin/hadoop-ec2 command (launch-cluster in particular). Please advise!

2012-04-05 17:06
by kee
Are you sure you want to setup a Hadoop cluster on EC2 instances manually? That's what Elastic Map Reduce is for - xxpor 2012-04-05 17:13
Yes at this point we want to evaluate both and decide - kee 2012-04-05 17:45


1

Specify your EC2 url in EC_URL environment variable like this:

export EC2_URL="https://ap-southeast-1.ec2.amazonaws.com/"

Then it will extract the info here and it is used by ec2-run-instances.

2012-04-06 01:13
by kee


1

If you start your instance by ec2-run-instances, you can specify --availability-zone. It is really a lot of work doing everything manually for hadoop. Do you want to try Apache Whirr?

2012-04-05 18:44
by root1982
It looks like the trend is using whirr? I will definitely try that. In the meantime I found an answer for my question - kee 2012-04-05 20:30
Ads