Redundancy and Data Protection

The redundancy level of a logical volume determines the number of copies of the parity bits that the Oracle FS System creates for the volume. The number of copies of the parity bits determines the level of protection for the volume data if a drive that contains any data for the volume fails.

Important! Contact Oracle Customer Support for assistance with sizing your system and with creating your logical volumes.

The redundancy Quality of Service (QoS) property provides the following choices for parity:

Single: Stores the original user data plus one set of parity bits to help in the recovery of lost data. Access to the data is preserved even after the failure of one drive. Single parity is implemented using RAID 5 technology and is the default redundancy level for the Storage Classes that specify the performance-type media.
Double: Stores the original user data plus two sets of parity bits to help in the recovery of lost data. Access to the data is preserved even after the simultaneous failure of two drives. Double parity is implemented using RAID 6 technology and is the default redundancy level for the Storage Classes that specify the capacity-type media.

A drive group is a logical object that manages a collection of drives of the same Storage Class, all of which are from the same Drive Enclosure. The number of drives in the drive group depends on the drive type. A drive group containing hard disk drives (HDDs) consists of 12 drives. A drive group containing solid state drives (SSDs) consists of six drives.

The following list shows the number of drive groups across which the Oracle FS System allocates capacity for a logical volume that is using regular QoS (a single-tiered volume). The number of drive groups depends only on the priority level that you have chosen for the volume:

Premium: 4 drive groups
High: 4 drive groups
Medium: 3 drive groups
Low: 2 drive groups
Archive: 2 drive groups

Note: For solid-state drives (SSDs), the system uses all of the assigned SSDs in the Storage Domain when striping a volume, regardless of the priority level chosen for the volume.

The number of drive groups that the Oracle FS System uses to stripe a single-tiered volume can, at times, be fewer than what is shown in the above list. The system might use fewer drive groups because, for example, many of the drive groups in the Storage Domain are nearing full capacity or are at full capacity. When either situation exists, the system might allocate more than one stripe for a new volume in the same drive group, degrading the overall performance for that volume.

When performance is degraded because the striping for the volume is across fewer drive groups than would normally be used, you can re-establish high performance for the volume by including additional drive groups in the Storage Domain.

Adding drive groups allows the Oracle FS System to rebalance the allocations so that the volume utilizes the proper number of drive groups. The rebalancing is achieved by the system moving the extra stripes from the drive group to the newly added drive group.