This is a distributed object storage system, aimming to realize Hadoop HDFS.
- Registration & Heartbeat Management of Datanode
- Block transfer between nodes in TCP
- Initial RPC interface design
- Datanode boot check & verfiy of existed blocks
- Simple snapshot of Block Tree in Namenode
- Block report construction
- Block recovery when one Datanode is down
- Multi nodes test.
This project is still in the beginning stage and welcome Pr or Fork.