Clustering concepts

Mirrored clusters

What mirrored clusters contain

A mirrored cluster contains two complete copies of the specified data volumes or file systems that you indicated as being mirrored volumes or file systems in your cluster. These copies are called plexes and are continuously and synchronously updated every time Data ONTAP writes data to the disks. Plexes are physically separated from each other across different groupings of disks.

Advantages of mirrored clusters

The advantages of having duplicate copies of your data are as follows:

Mirrored data on your cluster survives a double disk error in one of your disk shelves.
Mirrored data on your cluster survives if one part of the cluster loses connectivity to your disks.
Mirrored data on your cluster survives if your cluster loses connectivity to remote storage.
Mirrored data on your cluster survives the destruction of a mirrored site or destruction of connectivity to the remote site.

In any of these situations, the surviving mirrored plex continues to serve data until the problem is corrected. See the Data Protection Guide for detailed information about disaster protection through mirroring.

How Data ONTAP works with mirrored clusters

During configuration of the cluster, Data ONTAP identifies spare disks and divides them into separate groupings called pools. These pools of disks are physically separated from each other, allowing for high availability of mirrored volumes. When you add a mirrored volume or add disks to one side of a mirrored volume, Data ONTAP determines how much storage you need for the second half of the mirror, and dedicates that storage from a separate pool to the mirrored volume.

Mirrored cluster setup requirements and restrictions

The following list specifies the requirements and restrictions you must be aware of when setting up a new mirrored cluster or converting an existing cluster to a mirrored cluster:

Disks in the same volume must be part of the same plex, with its mirror in a different plex.
Disk volumes that you want to mirror must have at least an equivalent number of disks in the mirrored plex of the same or greater disk size.
Disk pool 0 volumes must be cabled to FC-AL or dual Fibre Channel adapters in node slots 2 through 5, if available.
Disk pool 1 volumes must be cabled to FC-AL or dual Fibre Channel adapters in node slots 8 through 11.
The two halves of the mirrored disk should have no common components.
Local SyncMirror must be licensed and enabled on the local and partner nodes.
There must be sufficient spares in each pool to account for a disk failure.
Cluster nodes that are separated by more than 30 meters must use an LC-LC 50/125 cable to connect their IB cluster interconnect adapters.
Clusters that are separated by more than 30 meters must use an SC-SC 50/125 plenum cable to connect the local node to the partner disk shelves.