While I was originally skeptical of transparent encryption, I like the value proposition of transparent encryption. HDFS has several layers, protocols and tools. While the HDFS core part seems to be well done in the Jira, inserting the matching transparency in the other tools or protocols need to be worked through.
I have the following areas of concern: - Common protocols like webhdfs should continue to work (the design doc marks this as a goal), This issue is being discussed in the Jira but it appears that webhdfs does not currently work with encrypted files: Andrew say that "Regarding webhdfs, it's not a recommended deployment" and that he will modify the documentation to match that. Aljeandro say "Both httpfs and webhdfs will work just fine" but then in the same paragraph says "this could fail some security audits". We need to resolve this quickly. Webhdfs is heavily used by many Hadoop users. - Common tools should like cp, distcp and HAR should continue to work with non-encrypted and encrypted files in an automatic fashion. This issue has been heavily discussed in the Jira and at the meeting. The /.reserved./.raw mechanism appears to be a step in the right direction for distcp and cp, however this work has not reached its conclusion in my opinion; Charles are I are going through the use cases and I think we are close to a clean solution for distcp and cp. HAR still needs a concrete proposal. - KMS scalability in medium to large clusters. This can perhaps be addressed by getting the keys ahead of time when a job is submitted. Without this the KMS will need to be as highly available and scalable as the NN. I think this is future implementation work but we need to at least determine if this is indeed possible in case we need to modify some of the APIs right now to support that. There are some other minor things under discussion, and I still need to go through the new APIs. Unfortunately at this stage I cannot give a +1 for this merge; I hope to change this in the next day or - I am working with the Jira's team. Alejandoro, Charles, Andrew, Atm, ... to resolve the above as quickly as possible. Sanjay (binding) On Aug 8, 2014, at 11:45 AM, Andrew Wang <andrew.w...@cloudera.com> wrote: > Hi all, > > I'd like to call a vote to merge the fs-encryption branch to trunk. > Development of this feature has been ongoing since March on HDFS-6134 and > HADOOP-10150, totally approximately 50 commits. > > ..... > Thanks, > Andrew -- CONFIDENTIALITY NOTICE NOTICE: This message is intended for the use of the individual or entity to which it is addressed and may contain information that is confidential, privileged and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient, you are hereby notified that any printing, copying, dissemination, distribution, disclosure or forwarding of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and delete it from your system. Thank You.
