RAID Recovery™
Recovers all types of corrupted RAID arrays
Recovers all types of corrupted RAID arrays
Last updated: Jan 27, 2025

RAID-Z and Its Difference Between RAID-Z2 vs RAID-Z3

Here you will find out:

  • what is RAID-Z
  • RAID-Z vs RAID-Z2 vs RAID-Z3

Are you ready? Let's read!

What is RAID-Z?

This is a non-standard RAID that uses the ZFS file system; no other file system can be used for this array. Note that there is not a single hardware controller that implements RAID-Z.

The ZFS file system uses an additional level of checksums to search for data corruption without displaying the appropriate messages. ZFS uses checksums with any level of redundancy, including single disk pools. The distribution mechanism is similar to RAID 5, but it uses dynamic bandwidth. Any block, regardless of its size, has its own RAID bandwidth, which means that each RAID-Z record is a full-band record.

In addition, this array has very high performance and speed, even better than RAID 5 in this regard.

Self-healing RAID-Z

The self-healing data function is very interesting and useful. It works as follows. When a RAID-Z block is read, the ZFS file system compares this block with the checksum. If the result of this comparison is unsatisfactory, then ZFS reads the parity information and checks which drive returned the wrong data.

After that, the damaged information is automatically restored and returned to the right place.

RAID-Z vs RAID-Z2 vs RAID-Z3

RAID-Z

  • Fault tolerance: Protects against the failure of 1 disk.
  • Parity scheme: Single parity (similar to RAID-5).
  • Data layout: ZFS arranges blocks across all drives (except one in each write group is used for parity) to provide redundancy.
  • Capacity overhead: The capacity of one drive’s worth of space is effectively consumed by parity. For example, if you have 5 drives (each 4 TB), you end up with about 16 TB usable capacity (4×4 TB) because 1 drive’s capacity is used for parity.
  • Performance considerations: Reads generally benefit from multiple disks reading in parallel. Writes can be slower than a pure stripe (RAID-0) because parity must be calculated and written. However, ZFS’s copy-on-write and intelligent caching help ameliorate overhead.
  • Use cases: Smaller (4–8 drive) arrays or situations where losing 2 disks at once is unlikely and cost efficiency is more important than very high redundancy.

RAID-Z2

  • Fault tolerance: Protects against the failure of 2 disks.
  • Parity scheme: Double parity (similar to RAID-6).
  • Data layout: ZFS writes two parity blocks for each stripe, allowing the pool to survive two simultaneous disk failures, or one failure plus an additional disk with uncorrectable errors.
  • Capacity overhead: The capacity of two drives’ worth of storage is reserved for parity. For instance, with 6 drives (each 4 TB), the total usable capacity is about 16 TB (4×4 TB) because 2 drives’ worth of capacity goes to parity.
  • Performance considerations: Slightly more overhead than RAID-Z in parity calculations, but readings still take advantage of parallel I/O.
  • Use cases: Medium to large ZFS pools (often more than 6 drives); recommended for scenarios where reliability is critical, such as small business servers or homelabs with valuable data.

RAID-Z3

  • Fault tolerance: Protects against the failure of 3 disks.
    Parity scheme: Triple parity.
  • Data layout: Three parity blocks are maintained, allowing the pool to survive up to three drive failures at once.
  • Capacity overhead: The capacity of three drives is consumed for parity. With 8 drives (each 4 TB), the total usable capacity is roughly 20 TB (5×4 TB).
  • Performance considerations: More CPU overhead during writes because of additional parity calculations, but modern systems often handle this without noticeable slowdown, especially if equipped with a decent CPU. Reads are still distributed, so read performance remains good.
  • Use cases: Large-scale deployments, enterprise environments, or extremely mission-critical data. For example, data centers with major reliability standards or archival libraries that can’t risk multiple failures.

Key considerations

  1. 1. Number of drives:
  • RAID-Z requires at least 3 drives.
  • RAID-Z2 requires at least 4 drives (though ZFS will let you create a RAID-Z2 with 3 drives, it’s impractical because overhead would be high).
  • RAID-Z3 typically starts to make sense at 5–6 drives or more.
  1. 2. Risk tolerance and data criticality:
  • RAID-Z (single parity) is often used for personal systems, home labs, or smaller setups.
  • RAID-Z2 (double parity) is for users or organizations with moderate to high reliability demands.
  • RAID-Z3 (triple parity) is generally for enterprise-grade or very large deployments, or those with extremely low tolerance for downtime or data loss.
  1. 3. Performance:
  • Read performance in all RAID-Z variants benefits from data being distributed across multiple disks.
  • Write performance can suffer as parity calculations increase with each additional parity block. However, ZFS’s copy-on-write design, intelligent caching (ARC, L2ARC), and optionally ZIL (ZFS Intent Log) devices can help mitigate performance bottlenecks.
  1. 4. Storage overhead:
  • RAID-Z sets aside one disk’s worth of space for parity (total of N-1 usable disks).
  • RAID-Z2 sets aside two disks’ worth of space for parity (N-2 usable disks).
  • RAID-Z3 sets aside three disks’ worth of space (N-3 usable disks).
  1. 5. ZFS features integration:
  • Checksumming ensures data integrity (each data block is checksummed, so silent corruption can be detected and repaired).
  • Snapshots, replication, and hybrid storage pools (using fast SSD-based devices for caching) are among the advanced ZFS features that combine well with RAID-Z.

Data recovery from RAID-Z

Disk failure can occur in several ways:

  • When a disk is completely damaged or you get messages that certain sectors are damaged
  • When data is corrupted without any warning

In the first case, RAID-Z can, when it is known which block or sector is damaged, independently reconstruct the data.

When a disk is completely damaged, this array behaves like a traditional one. That is, it is clear that all sectors are damaged and should be restored.

When dealing with noiseless data corruption, RAID-Z can help you recover it with the extra checksum provided by ZFS.

If automatic data recovery has not occurred or is simply not possible, you should consider a special RAID recovery application. Users of DiskInternals RAID Recovery application unanimously say that this is a great software solution, since the recovery results are stunning!

Tip: RAID 0 failed!

Ready to get your data back?

To start recovering your data, documents, databases, images, videos, and other files from your RAID 0, RAID 1, 0+1, 1+0, 1E, RAID 4, RAID 5, 50, 5EE, 5R, RAID 6, RAID 60, RAIDZ, RAIDZ2, and JBOD, press the FREE DOWNLOAD button to get the latest version of DiskInternals RAID Recovery® and begin the step-by-step recovery process. You can preview all recovered files absolutely for free. To check the current prices, please press the Get Prices button. If you need any assistance, please feel free to contact Technical Support. The team is here to help you get your data back!

The utility supports both manual and fully automatic determination of the main parameters of the array (array type, RAID controller type, strip size and disk order, etc.). It then automatically restores the data. All this is possible thanks to the Recovery Wizard, who will support you at each step of the process. In addition, the preview of the results will please everyone, since it is free of charge and only after you get it do you need to decide whether to purchase a license or not. Learn more about RAID 2 vs RAID 3.

Note: RAID 5 rebuild time.

Good luck!

Related articles

FREE DOWNLOADVer 6.23, WinBUY NOWFrom $249

Please rate this article.
4.516 reviews