You need to mount the EBS volume as a local filesystem and configure hadoop to use the filesystem mount point as its data path.
On Thu, 2009-01-29 at 19:49 -0500, FRANK MASHRAQI wrote: > Hello, > > I am trying to setup a Hadoop cluster with EBS. I cannot find where/how to > specify which EBS volumes to use for master/slaves in the scripts that are > in hadoop-0.19.0/src/contrib/ec2 directory. I am able to setup the hadoop > cluster manually but would like to automate launch and termination of > cluster. It seems the scripts assume that one will use S3 for Hadoop. Can > anyone guide me on how to pass EBS volumes information to an instance when > it is launched? > > Thanks, > Frank