Hi HDFS dev,
We would like to propose merging the development branch HDFS-6584 back to trunk. The work of HDFS-6584 is to support Archival Storage in HDFS. Hadoop needs a solution to decouple growing storage capacity from compute capacity. Nodes with higher density and less expensive storage with low compute power are becoming available and can be used as cold storage in the clusters. Based on policy, the data from hot can be moved to the cold. Adding more nodes to the cold storage can grow the storage independent of the compute capacity in the cluster. The Archival Storage feature introduces a new concept of storage policies (Hot, Warm, Cold) to HDFS and allows files to be stored in different storage medias (DISK, ARCHIVE) according to the storage policy. A new data migration tool is added for archiving data. For more details, please see the design doc posted in HDFS-6584. There are currently 22 sub-tasks in HDFS-6584. The main feature of HDFS-6584 is implemented and tested. The remaining work is adding user documentation, adding a test plan doc, adding more tests and fixing some minor bugs. We will finish all of the remaining work before merging the branch to trunk. The contributors of this work include Jing Zhao, Arpit Agarwal, Vinayakumar B, Suresh Srinivas, Sanjay Radia, Jitendra Pandey, Haohui Mai and Tsz-Wo Nicholas Sze. We thank the community for giving valuable comments and feedbacks on the design of this work. The vote will last for 7 days. Of course, I am +1 on it. Regards, Tsz-Wo