As support for AWS S3 is advertised and seems to be a common way to try Drill with existing data, it would be nice to have S3 support fully built-in. Having to search for and follow multi-step documentation on a blog post, to enable Jets3t including a protocol scheme error corrected only by a third-party comment was not a good experience. I have scripted this, you can find it at https://github.com/awslabs/emr-bootstrap-actions/pull/105
Additionally, Jets3t seems to have run its course; the prevailing advice in the community is to switch to the AWS SDK at this point. One advantage of doing so would be to support the EC2 instance profile credential mechanism, to let people avoid specifying credentials in configuration. Jets3t has stated that they are not interested in supporting this. Frustratingly, Drill (or maybe it¹s Jets3t) seems to have an interpretation of the various s3 protocol schemes that differs from Hadoop¹s (http://wiki.apache.org/hadoop/AmazonS3), yet this Hadoop page has been referenced in Drill configuration advice (I think it was on this mailing list). Paul
