null

RAIDZ1 vs RAIDZ2

Enterprise Resource Center • ZFS Knowledge Base

RAIDZ1 vs RAIDZ2

RAIDZ1 and RAIDZ2 are two of the most common parity-based VDEV layouts in ZFS. RAIDZ1 favors greater usable capacity with single parity, while RAIDZ2 sacrifices additional capacity to provide double-parity protection.

RAIDZ1 RAIDZ2 Single vs Double Parity ZFS VDEV Design
Tech Supply Direct
Quick Answer

What Is the Difference Between RAIDZ1 and RAIDZ2?

RAIDZ1 uses single parity and can tolerate one member-device failure, while RAIDZ2 uses double parity and can tolerate two member-device failures within the same VDEV.

RAIDZ1 generally provides more usable capacity from the same number of equal-size drives. RAIDZ2 uses additional capacity for parity in exchange for substantially greater protection against multiple device failures.

At a Glance

RAIDZ1 vs RAIDZ2 Comparison

Characteristic RAIDZ1 RAIDZ2
Parity Single parity Double parity
Device Failures Tolerated 1 per VDEV 2 per VDEV
Technical Minimum 2 devices 3 devices
OpenZFS Recommended Minimum 3+ devices 5+ devices
Simplified Raw Capacity (N − 1) × drive size (N − 2) × drive size
Capacity Efficiency Higher Lower
Fault Tolerance Lower Higher
After One Device Failure No additional device-failure tolerance remains Can still tolerate one additional device failure
Primary Tradeoff Capacity efficiency Additional redundancy
Single Parity

RAIDZ1

RAIDZ1 provides one parity level. It prioritizes capacity efficiency while still protecting the VDEV against one member-device failure.

Failure Tolerance 1 device
Simplified Capacity (N − 1) × drive size
Main Advantage Higher usable capacity
Main Tradeoff Only one-device fault tolerance
Double Parity

RAIDZ2

RAIDZ2 provides two parity levels. It gives up additional raw capacity in exchange for protection against two member-device failures within the VDEV.

Failure Tolerance 2 devices
Simplified Capacity (N − 2) × drive size
Main Advantage Greater fault tolerance
Main Tradeoff Additional parity capacity
Capacity Comparison

RAIDZ1 Provides More Usable Capacity

With the same number of equal-size drives, RAIDZ1 provides approximately one additional drive's worth of raw usable capacity compared with RAIDZ2 because RAIDZ2 dedicates an additional parity level.

RAIDZ1
(N − 1) × Drive Size

One parity level

RAIDZ2
(N − 2) × Drive Size

Two parity levels

These formulas are simplified raw-capacity estimates. Actual usable ZFS capacity depends on allocation behavior, metadata, block sizes, sector sizes, units and operational free space.

Example

6 × 12 TB Drives: RAIDZ1 vs RAIDZ2

Consider six equal-size 12 TB drives providing 72 TB of total raw device capacity.

Configuration Raw Capacity Parity Equivalent Simplified Usable Raw Capacity Failures Tolerated
RAIDZ1 72 TB ≈ 12 TB ≈ 60 TB 1
RAIDZ2 72 TB ≈ 24 TB ≈ 48 TB 2

The example is intentionally simplified. Formatted and practically usable capacity will be lower than the raw decimal TB figures shown.

Fault Tolerance

RAIDZ2 Provides an Additional Layer of Device-Failure Protection

RAIDZ1 can tolerate one unavailable member device. If another device in the same VDEV becomes unavailable before redundancy is restored, RAIDZ1 has exceeded its parity tolerance.

RAIDZ2 can tolerate two unavailable member devices. After the first device failure, one additional parity level remains, providing protection against another device failure while the VDEV is degraded.

RAIDZ1 After One Failure The VDEV remains accessible, but its single-device failure tolerance has been consumed.
RAIDZ2 After One Failure The VDEV remains accessible and can still tolerate one additional member-device failure.
Recovery Window

RAIDZ1 vs RAIDZ2 During Resilvering

The difference in parity becomes particularly important after a device fails. Replacing a failed device requires ZFS to restore the necessary data to the replacement device through resilvering.

A RAIDZ1 VDEV with one failed device has no remaining device-failure tolerance during this recovery period. A RAIDZ2 VDEV with one failed device still has one parity level available and can therefore survive another member-device failure.

This distinction becomes increasingly important when using large-capacity drives, when substantial amounts of data are allocated, or when replacement hardware may not be immediately available.

Large-Capacity Storage

Does Drive Size Affect the RAIDZ1 vs RAIDZ2 Decision?

Drive capacity does not change the number of failures each RAIDZ level can tolerate, but larger drives can increase the amount of data involved in replacement and recovery.

Larger Capacity Larger drives can hold substantially more allocated data that may need to be reconstructed after a failure.
Recovery Duration Recovery time depends on allocated data, device performance, system load and topology.
Risk Tolerance Additional parity may be more valuable when the consequences of another device failure are significant.
Replacement Availability Environments where replacement drives may be delayed can benefit from greater failure tolerance.
Performance

RAIDZ1 vs RAIDZ2 Performance

RAIDZ1 and RAIDZ2 performance cannot be compared accurately from parity level alone. VDEV width, block size, recordsize, storage media, workload and the number of top-level VDEVs all influence actual performance.

Write Workloads RAIDZ writes require data and parity operations across the participating columns of the stripe.
Random I/O Random-I/O-intensive workloads may justify evaluating mirrored VDEVs rather than selecting solely between RAIDZ1 and RAIDZ2.
Sequential I/O RAIDZ layouts can be well suited to capacity-oriented sequential workloads.
Pool Topology Multiple top-level VDEVs can provide more aggregate I/O capability than a single VDEV.
Capacity Priority

When RAIDZ1 May Make Sense

RAIDZ1 may be appropriate when capacity efficiency is important and the workload can accept single-device failure tolerance.

  • Smaller VDEV configurations
  • Less critical data
  • Capacity-sensitive deployments
  • Systems with strong independent backups
  • Environments with rapid drive replacement
Redundancy Priority

When RAIDZ2 May Make Sense

RAIDZ2 may be preferable when additional device-failure protection is worth sacrificing another drive's equivalent capacity.

  • Important production data
  • Large-capacity HDD deployments
  • Larger VDEVs
  • Longer potential recovery windows
  • Systems where replacement may be delayed
  • Environments requiring greater fault tolerance
Choosing a Layout

Should You Choose RAIDZ1 or RAIDZ2?

The decision is primarily a tradeoff between usable capacity and fault tolerance. RAIDZ1 gives more usable capacity from the same drive count, while RAIDZ2 provides another layer of protection against device failure.

Choose RAIDZ1 When... Capacity efficiency has greater value and one-device failure tolerance satisfies the system's risk requirements.
Choose RAIDZ2 When... Protecting against a second device failure is more important than retaining the capacity consumed by the additional parity level.
More Redundancy

What About RAIDZ3?

RAIDZ3 adds a third parity level and can tolerate three unavailable member devices within the VDEV. It sacrifices additional capacity in exchange for greater fault tolerance.

RAIDZ3 may warrant consideration for high-capacity storage systems where the additional protection is more valuable than maximizing usable capacity.

Data Protection

Neither RAIDZ1 nor RAIDZ2 Is a Backup

RAIDZ redundancy protects against certain storage-device failures. It does not independently protect against accidental deletion, administrative mistakes, malware, application-level corruption, catastrophic hardware loss or site-level events.

Important data should still be protected by an independent backup strategy appropriate for the required recovery objectives.

Frequently Asked Questions

RAIDZ1 vs RAIDZ2 FAQ

What is the main difference between RAIDZ1 and RAIDZ2?

RAIDZ1 uses single parity and tolerates one device failure. RAIDZ2 uses double parity and tolerates two.

Does RAIDZ1 have more usable capacity?

Yes. With the same number of equal-size devices, RAIDZ1 uses one parity level while RAIDZ2 uses two.

Is RAIDZ2 safer than RAIDZ1?

RAIDZ2 provides greater device-failure tolerance because it can withstand two member-device failures instead of one.

How many drives can RAIDZ1 lose?

A RAIDZ1 VDEV can tolerate one unavailable member device.

How many drives can RAIDZ2 lose?

A RAIDZ2 VDEV can tolerate two unavailable member devices.

Which is better for large hard drives?

Larger drives can increase the amount of data involved in recovery. RAIDZ2's additional parity may therefore be valuable when greater fault tolerance is required.

Is RAIDZ2 always better than RAIDZ1?

No. RAIDZ2 provides greater redundancy, but RAIDZ1 provides greater capacity efficiency. The appropriate choice depends on workload, topology, risk tolerance and recovery requirements.

Which is better during a resilver?

After one failure, RAIDZ2 retains tolerance for another member-device failure while RAIDZ1 has consumed its single parity level.

Is RAIDZ1 or RAIDZ2 a backup?

Neither. Both provide storage redundancy and should be combined with an independent backup strategy for important data.

Enterprise Storage Hardware

Designing a ZFS Storage System?

Tech Supply Direct can help identify compatible enterprise hard drives, SSDs, memory and server hardware for ZFS storage systems based on your platform, workload, capacity and redundancy requirements.