Table of ContentsView in Frames

Determining and enabling volume write-caching eligibility

Understanding whether the FlexVol volumes associated with an aggregate are eligible for write caching can help you ensure that the volumes with high performance requirements can get the maximum performance improvement from having their associated aggregate converted to a Flash Pool aggregate.

About this task

You can use this procedure to determine write caching eligibility to help you decide which aggregates are good candidates to become Flash Pool aggregates. You do not need any SSDs to complete this procedure.

Flash Pool aggregates employ two types of caching: read caching and write caching. Read caching is available for all volumes. Write caching is usually available for most volumes, but might be disabled due to the following reasons:

If the aggregate you want to convert to a Flash Pool aggregate was created using Data ONTAP 7.1 or earlier and you want write caching to be enabled for its associated volumes, you must copy each FlexVol volume associated with the aggregate to a new aggregate and convert the new aggregate to a Flash Pool aggregate. If some volumes are ineligible due to an ID collision, you can resolve the ID collision by moving the ineligible volumes to another aggregate (and back again if needed) or simply proceed with the conversion, using the force option. In this case, the ineligible volumes do not benefit from the Flash Pool aggregate cache for write operations.

Steps

  1. Attempt to enable the Flash Pool aggregate capability on the aggregate: aggr options aggr_name hybrid_enabled true
  2. Take the applicable action based on the result of Step 1:
    If... Then...
    The Flash Pool aggregate capability is successfully enabled Disable the Flash Pool aggregate capability again: aggr options aggr_name hybrid_enabled false

    You have completed this task. All of the volumes associated with the aggregate are eligible for write caching.

    The command displays an error message telling you that the aggregate cannot be converted to a Flash Pool aggregate Go to the next step.
  3. Take the applicable action based on why the volumes are ineligible and your requirements for those volumes:
    If the volumes are ineligible because... Then...
    The aggregate was created using Data ONTAP 7.1 or earlier, and you do not need to enable write caching for its associated volumes You can convert the aggregate by using the force option.
    The aggregate was created using Data ONTAP 7.1 or earlier, and you want to enable write caching for its associated volumes Copy all of the volumes associated with this aggregate to a new aggregate, and then repeat this procedure on the new aggregate.
    There are ID collisions for some volumes in the aggregate, and you do not need to enable write caching for those volumes You can convert the aggregate by using the force option.
    There are ID collisions for some volumes in the aggregate, and you need to enable write caching for those volumes Complete the following steps:
    1. Determine which volumes have ID collisions by running the vol status -v command for each volume, and noting the ones that do not show read-write eligibility.
    2. Copy or move all but one of each set of volumes with the same conflicting ID to a different aggregate, and then move them back until no more volumes show an ID conflict.

      If the volumes are accessed using only SAN protocols, you can use DataMotion for Volumes to move the volumes nondisruptively. For more information, see the Data ONTAP SAN Administration Guide for 7-Mode.

Example with ID collision

The following example shows the system output when volumes within the aggregate have ID collisions:

sys1> aggr options aggr1 hybrid_enabled on
aggr options: Failed to set caching policies on Aggregate "aggr1" because one or more volumes 
in the aggregate is not eligible for write caching.

sys1> vol status -v vol8
...........

                                    svo_allow_rman=off, svo_reject_errors=off,
                                    no_i2p=off, fractional_reserve=100, extent=off,
                                    try_first=volume_grow, read_realloc=off,
                                    snapshot_clone_dependency=off,
                                    no_delete_log=off, dlog_hole_reserve=off,
                                    nbu_archival_snap=off
                Volume has clones: vol8_clone
                         Volume UUID: 4a7afb51-47cd-11e1-bebd-123478563412
                Containing aggregate: 'aggr1'
                        Volinfo mode: 7-mode
                Volume Instance UUID: 4a7afb94-47cd-11e1-bebd-123478563412
              Volume Provenance UUID: 4a7afb94-47cd-11e1-bebd-123478563412

                Plex /aggr1/plex0: online, normal, active
                    RAID group /aggr1/plex0/rg0: normal, block checksums
                    RAID group /aggr1/plex0/rg1: normal, block checksums

        Snapshot autodelete settings for vol8:
                                        state=off
                                        commitment=try
                                        trigger=volume
                                        target_free_space=20%
                                        delete_order=oldest_first
                                        defer_delete=user_created
                                        prefix=(not specified)
                                        destroy_list=none
        Volume autosize settings:
                               state=off
        Hybrid Cache:
                Eligibility=read
                Write caching ineligibility reason=ID Collision(25443)