Right now Whirr can only deploy clusters with a single type of instances as specified by whirr.hardware-id.
After a cluster is started Whirr allows a user to either run scripts on the remote nodes or destroy the cluster. We are working on adding the ability to resize running clusters: https://issues.apache.org/jira/browse/WHIRR-214 I encourage you to take the lead on adding a feature that allows a user to specify a different instance type for each template group. I think we need a new property like this: whirr.hardware-id.hadoop-namenode+hadoop-jobtracker=m2.xlarge Start from ClusterSpec.java -- Andrei On Thu, Dec 8, 2011 at 6:14 AM, Periya.Data <[email protected]> wrote: > Hi All, > Is there a way in Whirr to increase the size of each node in the > cluster...say from t1.micro to m1.small? Or better yet, selectively > increase sizes of nodes. > > Thanks, > PD. >
