The following documents started as a technical explanation of the terms "copy on write" as it applied to ZFS in NexentaStor 3 (and OpenZFS in NexentaStor 4) back in 2012. While a number of things have changed with NexentaStor 5 (particularly in the area of snapshot management and long-distance replication with the introduction of in-kernel High-Performance Replication), the foundational concepts remain relevant and provide useful insights into the core benefits of using OpenZFS as a core data management building block.
We recommend reading the following documents in order:
1. NexentaStor: ZFS Copy-on-Write, Checksums, and Consistency
This first paper in a series starts with three necessary concepts, and the subsequent papers build upon this beginning.
2. NexentaStor: ZFS Snapshots and Replication
This second paper builds upon the first by describing how ZFS handles snapshots and replication differently from most traditional storage vendors.
Read first: NexentaStor: ZFS Copy-on-Write, Checksums, and Consistency
3. NexentaStor: ZFS Initialization and Resilvering
This third paper builds upon the first two by describing how ZFS handles initialization of RAID sets and resilvering after disk failures differently from most traditional storage vendors.
Read first: NexentaStor: ZFS Snapshots and Replication
4. NexentaStor: An Introduction to ZFS's Hybrid Storage Pool
This fourth paper builds upon the first two by describing how ZFS handles hybrid storage infrastructures differently from most traditional storage vendors.
Read first: NexentaStor: ZFS Initialization and Resilvering
5. NexentaStor: Auto-Tiered Storage
ZFS hybrid storage pools intelligently combine DRAM, flash, and hard disk drives to achieve the right balance of cost and performance for any given working set, while reducing the need for administrators to constantly monitor storage for I/O bottlenecks.
Read first: NexentaStor: ZFS Initialization and Resilvering