Understanding Fabric Attached MetroClusters

Configuration overview

The cluster configuration for distances between 500 meters and 30 kilometers uses four Brocade switches in a fabric configuration to connect the two cluster nodes.

Definition of a switch fabric

A Fabric Attached MetroCluster contains two switch fabrics. A switch fabric consists of a switch on the local half of the cluster connected to a switch on the remote half of the cluster. The two switches are connected to each other through long distance ISL (Inter Switch Link) cables.

For example, the first fabric in a MetroCluster begins from Switch 1 on the local side and is completed by connecting the ISL cable to the first switch on the remote side--Switch 3. The second fabric is created using Switch 2 on the local side, connected through a second ISL cable, to the second switch on the remote side--Switch 4.

Note
Only one ISL is supported per fabric.

How switches are divided

The Fibre Channel switches are divided into banks and quadrants. Each switch has two banks and four quadrants, dividing the switch into equal parts. For example, a Brocade® 3800 switch has 16 ports, divided into two banks of eight ports per bank.

The banks are further subdivided into two quadrants of four ports each (two for the Brocade 3200 and 3250). These quadrants are also identified as pools, which function the same in a Fabric Attached MetroCluster as in a mirrored cluster disk pool. Quadrants 1 and 3 belong to pool 0, and quadrants 2 and 4 belong to pool 1. Each switch bank contains a pool 0 and a pool 1.

The following illustrations show a 16-port and an 8-port Brocade switch, divided into banks, quadrants, and pools.

Note
Pools can only be pool 0 or pool 1. However, Data ONTAP can assign arbitrary numbers to plexes. For simplicity, plexes are shown in this document as plex0 and plex1, but they could have different numbers in your installation.

Advantages of Fabric Attached MetroClusters

Fabric Attached MetroClusters provide the same advantages as nonswitched MetroClusters by mirroring the data in separate locations. They also provide the following additional advantages:

Expanded distance between halves of the cluster beyond 500 meters provides increased disaster protection.
Disk shelves and nodes are not connected directly to each other, but are connected to a fabric with multiple data routes, ensuring no single point of failure.

How Data ONTAP works with Fabric Attached MetroClusters

Data ONTAP functions the same as it does in a nonswitched MetroCluster. For more information, see How Data ONTAP works with nonswitched MetroClusters.

Setup requirements and restrictions

The setup requirements for a Fabric Attached MetroCluster are the same as for nonswitched MetroClusters, with the following exceptions:

Node requirements
- The nodes must be an F825 or F880 cluster pair, or a FAS900 or FAS3000 series cluster pair, configured for mirrored volume use.
- Each node must have one or two supported HBA kits installed and running. For Fabric Attached MetroClusters, only the QLogic 2342 (X2050A) FC HBA or the FAS3000 on-board ports are supported.
- Each node requires a QLA2352 VI-MC cluster adapter; the slot position is dependent on the type of storage system.
Disk and shelf requirements
- Only DS14 and DS14Mk2 disk shelves are supported.
- Only FC-type disks are supported; you cannot use ATA drives for Fabric Attached MetroCluster configurations.
- You can connect a maximum of 336 drives to a Fabric Attached MetroCluster.
  Note
  You can not exceed the maximum storage capacity for each node, regardless of the number of disks. For information about the maximum storage capacity for storage systems, see the System Configuration Guide on the NOW site.
Switch requirements
- Each pair of nodes must use two switches to communicate with the clients.
- Switches must be either Brocade® 3200, 3250, 3800, or 3850.
- Mixed switch types can be used, but must be the same type on each side of the cluster.
- Switches must be running the correct firmware version.
- Switches must have long distance Small Form-factor Pluggable (SFP) transceivers.
  Note
  You can download Brocade switch firmware from the NOW site at now.netapp.com.
License requirements
- cluster--Enables the two nodes to communicate and take over each other's functions if a failure occurs.
- syncmirror_local--Enables mirrored volumes.
- cluster_remote--Enables mirrored volumes over long distances and enables the local node to take over the functions of the remote node if a failure occurs.

Configuration limitations

The Fabric Attached MetroCluster configuration has the following limitations:

You cannot use the switches in a MetroCluster configuration to connect Fibre Channel tape devices, or for FCP traffic of any kind. You can connect only storage systems and disk shelves to the switches in a MetroCluster configuration.
You can connect a tape storage area network (SAN) to either of the nodes in a MetroCluster configuration, but the tape SAN must not use the switches used in the MetroCluster configuration.