What are the rules for Advanced Disk Partitioning?

Applies to

• ONTAP 9
• ADPv1 or ADPv2
• Root-data partitioning

Answer

The following rules are enforced automatically when Advanced Disk Partitioning for versions 1 & 2 as specified.

• ADPv1: 1 data partition (P1) and 1 root partition (P2)
• ADPv2: 2 equally sized data partitions (P1, P2) and 1 root partition (P3)
• ADPv1 is only supported for HDD and ADPv2 is only supported for SSD disks.
• Root partitions can only be used for node root aggregates. 
• A RAID group cannot have more than one partition from the same disk (ADPv2).
  • Example: 
    • For partitioned disk 0a.00.1, partitions 0a.00.1P1 and 0a.00.P2 can not be in the same RAID group. 
    • They can be in different RAID groups of the same aggregate.
• Partitioned and non-partitioned disks can be present together in the same aggregate, however, they must be in different RAID groups.
• The only supported RAID types for aggregates containing root-data partitioned disks (ADPv1) are RAID-TEC and RAID-DP.  RAID4 is not supported.
• The supported RAID types for aggregates containing root-data-data partitioned disk (ADPv2) are RAID-TEC, RAID-DP, and RAID4 (ONTAP 9.5 and later).
• All-Flash FAS (AFF) systems only support ADPv2.
• For ADPv1, the spares requirements are as shown:
  • There must be (1) whole HDD available on each node of the HA pair.  The disk can be partitioned, but both partitions must be available as spares. 
  • If there is a mix of partitioned disk and non-partitioned disk RAID groups, then the requirement are (2) whole spares for each node that has both a partitioned disk RAID group and a non-partitioned RAID group.
• In default for ADPv1, the owner of partitions are same as the container; for ADPv2, the owner of the secondary data partition should be different with the root partition and primary data partition.
• For AFF systems using ADPv2 require spares as shown
  • If all RAID groups are partitioned disks, with the same disk container type; (1) a whole spare per HA is required.  The spare can be partitioned across the nodes.
  • If there is a mix of partitioned disk and non-partitioned disk RAID groups, then the requirement is (1) a whole spare per node so that a spare is always available both for the partitioned disk RAID groups and the whole disk RAID groups.
• ADPv1 initializes an HA pair with up to 24 HDD partitions; however, more partitions can be created after initialization
• ADPv2 initializes an HA pair with up to 48 SSD partitions; however, more partitions can be created after initialization
• For ADPv2, the minimum recommended same size or type to satisfy the hot spares/node requirements within ONTAP with mixed whole and partitioned RAID groups:

Additional Information

• The following example error message is reported when attempting to change the RAID type to an unsupported version:

cluster1::> aggr modify -aggregate aggr0 -raidtype raid4

Error: command failed: Failed to modify raidtype of aggregate "aggr0".
Reason: Change to raidtype RAID4 is not supported. Fields that were successfully modified: none

• NetApp Hardware Universe contains extensive information for each supported platform on initial partition sizes for each supported disk type.