Understanding mirrored clusters

What mirrored clusters do

Mirrored clusters provide high availability through failover, just as standard clusters do. Additionally, mirrored clusters provide the capability to have two complete copies of all mirrored data. These copies are called plexes and are continually and synchronously updated every time Data ONTAP writes a mirrored aggregate. The plexes can be physically separated to protect against the loss of one set of disks.

Note
Mirrored clusters do not provide the capability to fail over to the partner node if one node is completely lost. For example, if power is lost to one entire node, including its storage, you cannot fail over to the other node. For this capability, use a MetroCluster.

Mirrored clusters use SyncMirror. For more information, see the Data Protection Online Backup and Recovery Guide.

The restrictions, requirements, and best practices for mirrored clusters are the same as for a standard cluster. For more information, see Failover event cause-and-effect table.

Advantages of mirrored clusters

Mirroring your data protects it from the following problems, both of which would cause data loss without mirroring:

The failure or loss of two or more disks in a RAID4 aggregate
The failure or loss of three or more disks in a RAID-DP^TM aggregate

In addition, the failure of an FC-AL adapter, loop, or ESH module does not require a cluster failover in a mirrored cluster.

Similar to standard clusters, if either node in a mirrored cluster becomes impaired or cannot access its data, the other node can automatically serve the impaired node's data until the problem is corrected.

Setup requirements and restrictions

The restrictions and requirements for mirrored clusters include those for a standard cluster, as outlined in Failover event cause-and-effect table. In addition, the following list of restrictions applies when setting up a mirrored cluster:

Disks in the same plex must be from the same pool, with disks in the opposite plex from the opposite pool.
If software disk ownership isn't used, you must observe the SyncMirror pool rules for direct-attached Fibre Channel disks.
For more information, see SyncMirror pool rules for direct-attached Fibre Channel disks.
There must be sufficient spares in each pool to account for a disk failure.
The following licenses must be enabled on the local and partner nodes:
- cluster
- syncmirror_local

SyncMirror pool rules for direct-attached Fibre Channel disks

The FAS3000 series storage system uses new slots for direct-attached Fibre Channel disks.

Note
This is valid only if software disk ownership is not being used.

Follow these SyncMirror pool rules for direct-attached Fibre Channel disks:

For FAS3000 series

pool 0 - Onboard adapters 0a, 0b, and adapters in PCI slots 1, 2
pool 1 - Onboard adapters 0c, 0d, and adapters in PCI slots 3, 4

For all other storage systems

pool 0 - Adapters in PCI slots 1 - 7
pool 1 - Adapters in PCI slots 8 - 11