About HDFS Snapshots
CDH 5 provides an HDFS snapshot capability. A snapshot is a copy of all or part of the file system at a given point in time. Important characteristics of HDFS snapshots include:
- You can take a snapshot of the entire file system, or of any directory once you have made the directory snapshottable (hdfs dfsadmin -allowSnapshot <path>).
- There is no limit on the number of snapshottable directories.
- You can keep up to 65,536 snapshots of each directory.
- After you have taken the first snapshot, subsequent snapshots consume only as much disk space as the delta between data already in snapshot and the current state of the live file system.
- Snapshots are instantaneous and do not interfere with other HDFS operations.
For more information about snapshots, and instructions for taking and maintaining them, see HDFS Snapshots.