Zpool wizard

A zpool is the foundational storage construct in ZFS. It serves as a logical storage pool that combines multiple physical storage devices (disks) into vdevs (virtual devices), which collectively form the unified zpool. From this zpool, ZFS creates and manages datasets (file systems) and zvols (block storage volumes).

The zpool wizard is made up of the following steps:

1. Add data group

This section provides information about all storage devices connected to the storage server. To add the first Data Group to your Zpool, follow these steps:

1. Select the desired disks from the list on the left.
2. Choose the redundancy type.
3. Click the "Add group" button.

The available redundancy options for groups are as follows:

• Single: Each disk operates as an independent drive with no redundancy.
• Mirror: All data written to one device in the mirror is automatically replicated to another device, ensuring data redundancy. A minimum of two disks is required to create a mirrored vdev.
  • Mirror (Single Group): All selected disks will be combined into a single mirrored group.
  • Mirror (Multiple Groups): The selected disks will be paired into multiple mirrored groups, each consisting of two disks.
• RAIDZ-1: Allows for the failure of one disk per RAIDZ-1 group without losing data. A minimum of three disks is required for a RAIDZ-1 group.
• RAIDZ-2: Allows for the failure of two disks per RAIDZ-2 group without losing data. A minimum of four disks is required for a RAIDZ-2 group.
• RAIDZ-3: Allows for the failure of three disks per RAIDZ-3 group without losing data. A minimum of five disks is required for a RAIDZ-3 group.

To learn more vdev types, please refer to the following article: Redundancy_in_Disks_Groups

2. Add write log

This feature allows you to configure the write log function using a chosen redundancy level (either a single drive or a mirror). The write log utilizes a separate intent log (SLOG) device. A fast SSD/NVME should be used for this vdev.

Key points to consider:

• If multiple log devices are specified, write operations are load-balanced between the devices.
• Log devices can be configured with redundancy by using mirrors to enhance fault tolerance.
• RAIDZ vdev types are not supported for the intent log.

This ensures efficient and reliable write operations while leveraging the selected redundancy level.

3. Add read cache

A cache device is used to store frequently accessed storage pool data, providing an additional layer of caching between main memory and disk. These devices cannot be configured as mirrors or RAIDZ groups. A fast SSD/NVME should be used for this vdev.

Key benefits and considerations:

• Cache devices are particularly useful for read-heavy workloads where the working dataset size exceeds the capacity of main memory.
• By utilizing cache devices, a larger portion of the working dataset can be served from low-latency storage, improving performance significantly.
• The greatest performance improvements are seen in workloads characterized by random reads of primarily static content.

Adding a read cache helps enhance performance and reduces latency for storage systems with high read demands.

4. Add special devices group

Special devices are used to store specific types of data, such as metadata or small files, on dedicated storage devices separate from the main data pool. A fast SSD/NVME should be used for this vdev.

Key features and benefits:

• Storing metadata on special devices improves performance for metadata-intensive operations, such as file lookups and directory traversals.
• Small files below a certain size threshold can also be stored on these devices, enhancing read and write speeds for such workloads.
• Special devices are particularly beneficial for environments with a large number of small files or high metadata activity.

Using special devices optimizes the overall performance of the ZFS pool by offloading critical metadata and small-file operations to faster storage.

5. Add deduplication group

A deduplication group can be explicitly excluded from a special device group as a dedicated storage group used to hold deduplication tables. This allows the deduplication tables to be stored separately from the special device class.

Key features and considerations:

• Storing deduplication tables in a dedicated group improves the efficiency of deduplication processes by isolating them from other metadata operations.
• This configuration provides flexibility in optimizing storage layout based on workload requirements.
• Using a deduplication group is particularly beneficial for systems with high deduplication demands, ensuring better performance and management.

This setup enhances deduplication performance while maintaining a clear separation of metadata and deduplication operations.

6. Add spare disks

A spare disk is a special pseudo-vdev used to track available spare devices for a zpool.

Using spare disks enhances the reliability of the storage pool by allowing seamless drive replacement and reducing the risk of data loss.

7. Configuration

During this step, you can configure the Zpool by naming it and enabling additional features if required.

Key configurations:

• Zpool Name: Assign a unique and descriptive name to the Zpool for easy identification.
• Enable AutoTRIM: If supported by your devices, enable the AutoTRIM feature to automatically reclaim unused space. AutoTRIM helps optimize the performance and lifespan of SSDs by informing them when blocks are no longer in use.
• Small blocks policy settings if a special device group has been configured in Step. When the small block size is set for the pool all datasets inherit this value by default. It can be changed for a particular dataset in its setting.

Proper configuration ensures that the Zpool is tailored to your needs and operates efficiently.

8. Summary

This step provides a summary of the zpool configuration, detailing the arrangement of disk groups and their roles within the pool. Click ‘Add zpool’ to create a zpool.