Understanding standard clusters

What standard clusters contain

A standard cluster contains two sets of Fibre Channel disk shelves, a local node, and a partner node (except the FAS270c, which does not require two sets of disk shelves). Both nodes must be the same model.

The nodes are connected to each other through an IB cluster adapter or NVRAM5 adapter, which allows one node to serve data to the disks of its failed partner node. Each node continually monitors its partner, mirroring the data for each other's NVRAM.

Benefits of clusters

Configuring storage systems in a cluster provides the following benefits:

Fault tolerance
When one node fails or becomes impaired, a takeover occurs, and the partner node continues to serve the failed node's data.
Nondisruptive software upgrades
When you halt one node and allow takeover, the partner node continues to serve data for the halted node while you upgrade the node you halted.

For more information about nondisruptive upgrades, see the Upgrade Guide.
Nondisruptive storage system and disk maintenance
When you halt one node and allow takeover, the partner node continues to serve data for the halted node while you replace or repair hardware in the node you halted.

For more information about nondisruptive storage system maintenance, see Changing Clustered Hardware Nondisruptively.

How Data ONTAP works with standard clusters

In a standard cluster, Data ONTAP functions so that each node monitors the functioning of its partner through a heartbeat signal sent between the nodes. Data from the NVRAM of one node is mirrored by its partner, and each node can take over the partner's disks if the partner fails. Also, the nodes synchronize each other's time. For more information, see Cluster Management.

Cluster diagram

The following figure shows a common configuration of a standard cluster.

Note
The preceding diagram assumes that each node owns the disks attached to its own shelves, connected using the FC-AL A loop. If you are using a FAS270c (which contains two controller units) or the SnapMover® software, you have software disk ownership. In these cases, you need to establish disk ownership explicitly.

For more information about software disk ownership, see the Disk chapter of the Storage Management Guide. For more information about SnapMover, see the MultiStore® Management Guide.

Cluster characteristics

Storage systems in a cluster have the following characteristics:

They are connected to each other through a cluster interconnect consisting of adapters and cable through which they do the following tasks:
- Continually check whether the other node is functioning
- Mirror log data for each other's NVRAM
- Synchronize each other's time
They use two or more FC-AL loops in which the following conditions apply:
- Each node manages its own disks.
- Each node in takeover mode manages its partner's disks.
  Note
  For the FAS 270c and systems using the SnapMover software, software disk ownership is in use, so disk ownership is established explicitly rather than by which shelf the disk is attached to.
  
  For more information about software disk ownership, see the chapter on disk management in the Storage Management Guide. For more information about SnapMover, see the Multistore Management Guide.
They own their spare disks and do not share them with the other node.
They each have two mailbox disks on the root volume (four if the root volume is mirrored using the SyncMirror® feature). The mailbox disks are used to do the following tasks:
- Maintain consistency between the pair
- Continually check whether the other node is running or whether it has performed a takeover
- Store cluster information that is not specific to any particular node
They can reside on the same NT domain or on different domains.

Standard cluster setup requirements and restrictions

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

Architecture compatibility
Both nodes must be the same type and model.
Storage system capacity
The number of disks must not exceed the maximum cluster capacity. To determine the maximum capacity, see the System Configuration Guide on the NetApp on the Web^TM (NOW) site at Service & Support > Access Product Documentation > System Configuration Guide.
Disks and disk shelf compatibility
The disk shelves must have two paths so that each node can attach to the shelf.
Cluster interconnect adapters and cables must be installed.
Nodes must be attached to the same network and the NICs must be configured correctly.
Cluster software must be licensed and enabled on both nodes.
The same system software, such as CIFS, NFS, or SyncMirror, must be licensed and enabled on both nodes.
Note
MultiStore software can be licensed on one node only; this does not prevent a successful takeover of the partner node.

Best practices when deploying clusters

You can do many things to ensure that clustered storage systems achieve maximum uptime:

Make sure that the clustered controller units and shelves are on different power supplies/grids, so that a single power outage does not affect both controller units and shelves.
Use VIFs (virtual interfaces) to provide redundancy and improve availability of network communication.
Follow the documented procedures in the Upgrade Guide when upgrading clustered storage systems.
Maintain consistent configuration between clustered nodes. An inconsistent clustered storage system configuration is often a culprit for failover problems.
Test the failover capability routinely (for example, during planned maintenance) to ensure proper clustered storage system configuration.
Make sure that clustered nodes have sufficient resources to adequately support workload during takeover mode.
Use the Cluster Configuration Checker to help ensure that failovers are successful.
For more information, see Adding disk shelves to a cluster.