Table of ContentsView in Frames

Creating a Flash Pool aggregate

You create a Flash Pool aggregate by enabling the feature on an existing 64-bit aggregate composed of HDD RAID groups, and then adding one or more SSD RAID groups to that aggregate. This results in two sets of RAID groups for that aggregate: SSD RAID groups (the SSD cache) and HDD RAID groups.

Before you begin

About this task

After you add an SSD cache to an aggregate to create a Flash Pool aggregate, you cannot remove the SSD cache to convert the aggregate back to its original configuration.

You can change the RAID group size of the SSD cache, but you cannot make this change until after SSDs have been added. After disks have been added to a RAID group, they cannot be removed. If you know that you want to use a different RAID group size than the default SSD RAID group size, you can add a small number SSDs at first. Then, after you update the RAID group size, you can add the rest of the SSDs.

By default, the RAID level of the SSD cache is the same as the RAID level of the HDD RAID groups. You can override this default selection by specifying the raidtype option when you add the first SSD RAID groups. Although the SSD cache is providing caching for the HDD RAID groups, the SSD cache is integral to the health of the aggregate as a whole. An SSD RAID group that experiences a failure that exceeds the RAID protection capability of the RAID level in use takes the aggregate offline. For this reason, it is a best practice to keep the RAID level of the SSD cache the same as that of the HDD RAID groups.

There are platform- and workload-specific best practices for Flash Pool SSD cache size and configuration. For information about these best practices, see Technical Report 4070: NetApp Flash Pool Design and Implementation Guide.

Steps

  1. Mark the aggregate as hybrid-enabled: storage aggregate modify -aggregate aggr_name -hybrid_enabled true
    If this step does not succeed, determine write-caching eligibility for the target aggregate.
  2. Add the SSDs to the aggregate by using the storage aggregate add-disks command.
    You can specify the SSDs by ID or by using the diskcount and disktype parameters. You do not need to specify a new RAID group; Data ONTAP automatically puts the SSDs into their own RAID group.

    If you plan to change the RAID group size for the SSD cache, you should add only a small number SSDs in this step. (You must add at least three.)

    If the HDDs and the SSDs do not have the same checksum type, or if the aggregate is a mixed-checksum aggregate, then you must use the checksumstyle parameter to specify the checksum type of the disks you are adding to the aggregate.

    You can specify a different RAID type for the SSD cache by using the raidtype option.

  3. If you want a different RAID group size for the SSD cache than for the HDD RAID groups, change the SSD RAID group size: storage aggregate modify -aggregate aggr_name -cache-raid-group-size size
  4. If you did not add all of the required SSDs in the previous step, add the rest of the SSDs by using the storage aggregate add-disks command again.