Learn About Remote Mirroring

The Remote Mirroring premium feature is used for online, real-time data replication between storage subsystems over a remote distance. The mirroring is managed by the storage subsystem controllers and is transparent to host machines and applications. You create one or more mirrored logical drive pairs that consist of a primary logical drive at the primary site and a secondary logical drive at a secondary, remote site. After you create the mirror relationship between the two logical drives, the controller owner of the primary logical drive copies all of the data from the primary logical drive to the secondary logical drive. This is called a full synchronization .

Note that since replication is managed on a per-logical drive basis, you can mirror individual logical drives in a primary storage subsystem to appropriate secondary logical drives in several different remote storage subsystems.

Disaster Recovery

The secondary, remote logical drive is unavailable to secondary host applications while mirroring is in progress. In the event of a disaster at the primary site, you can fail over to the secondary site by performing a role reversal to promote the secondary logical drive to a primary logical drive. Then the recovery host will be able to access the newly promoted logical drive and business operations can continue.

Data Replication

When a primary controller (the controller owner of the primary logical drive) receives a write request from a host, the controller first logs information about the write to a special logical drive called a mirror repository logical drive . It then writes the data to the primary logical drive. The controller then initiates a remote write operation to copy the affected data blocks to the secondary logical drive at the remote site. After the host write to the primary logical drive is complete and the data has been copied to the secondary logical drive at the remote site, then the controller removes the log record on the mirror repository logical drive. Finally, the controller sends an I/O completion indication back to the host system. Because the controller does not send the I/O completion to the host until the data has been copied to both the primary and secondary logical drives, this mirroring operation is called synchronous.

When write caching is enabled on either the primary or secondary logical drive, the I/O completion is sent when data is in the cache on the side (primary or secondary) where write caching is enabled. When write caching is disabled on either the primary or secondary logical drive, then the I/O completion is not sent until the data has been stored to physical media on that side.

When a controller receives a read request from a host system, the read request is handled normally and no communication takes place between the primary and secondary storage subsystems.

Link Interruptions or Secondary Logical Drive Errors

Sometimes a primary controller receives a write request from a host that it can write to the primary logical drive, but a link interruption prevents communication with the secondary controller. In this case the remote write cannot complete to the secondary logical drive and the primary and secondary logical drives are no longer appropriately mirrored. The primary controller transitions the mirrored pair into Unsynchronized state and sends an I/O completion to the primary host. The primary host can continue to write to the primary logical drive but remote writes will not take place.

When connectivity is restored between the controller owner of the primary logical drive and the controller owner of the secondary logical drive, a full synchronization takes place. The mirrored pair transitions from an Unsynchronized state to a Synchronization in Progress state .

The primary controller also marks the mirrored pair as Unsynchronized when a logical drive error on the secondary side prevents the remote write from completing. For example, an offline or a failed secondary logical drive can cause the Remote Mirror to become Unsynchronized. When the logical drive error is corrected (the secondary logical drive is placed online or recovered to an Optimal state ) then a full synchronization automatically begins and the mirrored pair transitions to a Synchronization in Progress state.

Connectivity and Logical Drive Ownership

A primary controller will only attempt to communicate with its matching controller in the secondary storage subsystem. For example, Controller A in the primary storage subsystem only attempts communication with Controller A in the secondary storage subsystem. The controller (A or B) that owns the primary logical drive determines the controller owner of the secondary logical drive. If the primary logical drive is owned by Controller A on the primary side, the secondary logical drive is therefore owned by Controller A on the secondary side. If primary Controller A cannot communicate with secondary Controller A, no controller ownership changes take place.

When an I/O path error causes a logical drive ownership change on the primary side, or if the storage administrator changes the controller owner of the primary logical drive, the next remote write processed will automatically trigger a matching ownership change on the secondary side. For example, if a primary logical drive is owned by Controller A and then you change the controller owner to Controller B, the next remote write will change the controller owner of the secondary logical drive from Controller A to Controller B. Because controller ownership changes on the secondary side are controlled by the primary side, they do not require any special intervention by the storage administrator.

Controller Resets and Storage Subsystem Power Cycles

Sometimes a remote write is interrupted by a controller reset or a storage subsystem power cycle before it can be written to the secondary logical drive. The storage subsystem controller does not need to perform a full synchronization of the mirrored logical drive pair in this case. A controller reset causes a controller ownership change on the primary side from the preferred controller owner to the alternate controller in the storage subsystem. When a remote write has been interrupted during a controller reset, the new controller owner on the primary side reads information stored in a log file in the preferred controller owner's mirror repository logical drive. The information is used to copy the affected data blocks from the primary logical drive to the secondary logical drive, eliminating the need for a full synchronization of the mirrored logical drives.

Activating the Remote Mirroring Feature

Like other premium features, the Remote Mirroring feature is enabled by purchasing a feature key file from your storage supplier. You must enable the feature on both primary and secondary storage subsystems.

Unlike other premium features, you must also activate the feature after you enable it, using the Activate Remote Mirroring Wizard in the Subsystem Management Window. Each controller in the storage subsystem must have its own mirror repository logical drive for logging write information to recover from controller resets and other temporary interruptions. The Activate Remote Mirroring Wizard guides you to specify the placement of the two mirror repository logical drives (on newly created or existing free capacity in the storage subsystem).

After you activate the feature, one Fibre Channel host side I/O port on each controller is solely dedicated to Remote Mirroring operations. No host-initiated I/O operations are accepted by the dedicated port. I/O requests received on this port are accepted only from remote controllers that are participating in Remote Mirroring operations with the controller.

Connectivity requirements

Dedicated Remote Mirroring ports must be attached to a Fibre Channel Fabric environment with support for the Directory Service and Name Service interfaces.

You can use a Fabric configuration that is dedicated solely to the Remote logical drive Mirroring ports on each controller. In this case, host systems can connect to the storage subsystems using Fabric, FC-AL, or Point-to-Point configurations that are totally independent of the dedicated Remote Mirroring fabric.

Alternatively, you can use a single Fibre Channel Fabric configuration for both the Remote Mirroring connectivity and for the host I/O paths to the controllers.

The maximum distance between primary and secondary sites is 10 km, using single mode Fiber and Optical Long-Wave Giga-bit Interface Converters (GBICs).

Restrictions

The following restrictions apply to mirrored logical drive candidates and storage subsystem mirroring: