Example: Creating a 10+2+2 EC volume using maprcli

This example shows the steps to use maprcli to create a 10+2+2 erasure coded (EC) volume in an HPE Ezmeral Data Fabric cluster.

Prerequisites

The erasure coding scheme in this example requires at least 14 worker nodes in the Data Fabric cluster. The erasure coding scheme, 10+2+2 requires 10 data nodes, two global parity nodes, and two local parity nodes.

About this task

This example shows the steps to use maprcli to create a volume in a Data Fabric cluster that uses 10+2+2 erasure coding (EC), which is an erasure coding scheme with local parity. This example includes creating custom tiering rules and offload schedules.

Procedure

  1. Create a schedule.

    Schedules specify recurring points in time at which certain actions are to occur. In the following example, an erasure coding schedule specifies how often data is automatically offloaded to a different storage tier. Data is offloaded according to the tier rules.

    For more information about schedules, see Creating a Schedule in the HPE Ezmeral Data Fabric documentation.

    In the following example,
    • The name of the EC schedule is: my_ec_schedule
    • The offloading frequency is: hourly
    • The id is a unique number that is not one of the default schedules (1through 4). 
    maprcli schedule create -schedule '{ "id":5, "name":"my_ec_schedule", "inuse":0, "rules":[ { "frequency":"hourly", "retain":"10y" } ] }'
  2. Create an erasure coded tier with default values.

    In the following example, the type parameter value, ectier specifies that an erasure coded tier is created.

    maprcli tier create -name mywarm_tier -type ectier
  3. Create a tier rule.

    Tier rules define the criteria for offloading data.

    In the following example, the rule my_rule specifies data that is owned by the mapr user and was last modified 10 minutes (600 seconds) ago.

    maprcli tier rule create -name my_rule -expr "(u:mapr & m:600s)"
  4. Create an erasure coded volume.

    The ecscheme parameter specifies the erasure coding scheme. In the following example, the erasure coding scheme is 10+2+2, which specifies an erasure coding scheme with 10 data nodes, two local parity nodes, and two global parity nodes.

    For more information about erasure coding schemes, see Erasure Coding Scheme for Data Protection and Recovery in the HPE Ezmeral Data Fabric documentation.

    In the following example, the volume is enabled for data tiering by setting the tieringenable parameter to true. The tieringenable parameter must be used instead of the ecenable parameter because a tier name is also specified.

    maprcli volume create -name my_volume -path "/my_vol" -tieringenable true -tiername mywarm_tier -ecscheme "10+2+2" -ectopology "/data" -tieringrule my_rule -offloadschedule 5
  5. (Optional) Force an offload to occur. 
    maprcli volume offload -name my_volume