• Replication workflow

    PDF

    Replication workflow

    Configure replication

    Replication can occur locally or between a local storage system and a remote storage system. Use the native Replication feature in Unisphere to configure replication for block or file storage resources. The system supports asynchronous replication of all storage resources and NAS servers. For block-based storage resources (LUNs, consistency groups, and VMware VMFS datastores), synchronous replication is also supported.

    You can also configure synchronous and asynchronous replication for LUNs and VMware VMFS datastores using the RecoverPoint appliance-based solution.

    Configure local replication

    Replication interfaces and connections do not apply to local replication. When using the UEMCLI or the REST API, once you create the identical source and destination storage resources or NAS servers on the storage system, you can proceed to configure a replication session. When using Unisphere, you only need to create the source storage resources or NAS servers on the storage system. Unisphere does not allow you to create a session with an existing destination. A DR_ is concatenated onto the resource name for local destinations to ensure that the source and destination names on the same system are unique (that is, LUN names need to be unique).

    Note the following:

    • For a disaster recovery scenario, it is recommended that the destination storage resource and NAS server are configured on a storage pool other than the pool used for the source storage resource and NAS server.
    • For a migration scenario, which means migrating the source storage resource and NAS server to a destination storage resource and NAS server on the same pool, use the CLI to configure local replication. The Unisphere GUI does not allow local replication between storage resources and NAS servers on the same pool.
    • When you configure local replication for a tenant, create a pool for the tenant on the destination system that matches the corresponding pool on the source system (if one exists). Then add the tenant to the destination system, using the same UUID and VLANs as the tenant on the source system.

    Configure asynchronous replication

    Before you begin

    If you are configuring asynchronous replication for a tenant, create a pool for the tenant on the destination system that matches the corresponding pool on the source system (if one exists). Then add the tenant to the destination system, using the same UUID and VLANs as the tenant on the source

    Procedure
    1. Configure the replication interfaces on each SP of the source and destination systems.
    2. Configure a replication connection using the Asynchronous connection mode.
    3. For file storage, create a replication session for the NAS server associated with the file storage.
      Storage resources included in a NAS server automatically get replicated when a replication session is first configured for the NAS server. The replication session for the storage resources will inherit the same attributes as the associated replication session of the associated NAS server. For the storage resources you do not want participating in replication, you can choose to remove the associated replication sessions manually.
    4. Create a remote replication session for the storage resource.
      You only need to configure replication interfaces and connections for the first replication session between two systems. The same connection can be used again for subsequent replication sessions between the same systems.

    Configure synchronous replication

    Synchronous replication is only supported for block storage.
    Procedure
    1. Identify the Synchronous Replication Fibre Channel (FC) ports on each system.
      To determine the FC port used for synchronous replication, in the CLI console, run the command /remote/sys show -detail. Port information, similar to the following example, will appear in the output:
                                    Synchronous FC ports = spb_fc4, spa_fc4
                                  

      For more information, see the Unisphere CLI User Guide.

    2. Zone the Synchronous Replication FC ports between the systems.
      If the source and destination systems are co-located, instead of zoning, you can choose to use direct-connected FC cables between the SPs.
    3. Configure the replication interfaces on each SP of the source and destination systems based on the connection mode you want:
      • For synchronous replication support (Synchronous connection mode), use the Synchronous Replication Management Ports on each SP of both the systems.
      • For asynchronous and synchronous replication support (Both connection mode), in addition to the replication interfaces for Synchronous Replication Management Ports, configure additional interfaces using the Ethernet Ports on each SP of the source and destination systems.
    4. Configure the replication connection between source and destination systems from the source system only.
      • For synchronous replication support, specify the Synchronous connection mode.
      • For asynchronous and synchronous replication support, specify the Both connection mode.
    5. Create the synchronous replication session.
      You only need to configure replication interfaces and connections for the first replication session between two systems. The same connection can be used again for subsequent replication sessions between the same systems.

    Replication interfaces

    Each SP must include at least one replication interface configured on it. Each replication interface on the SP of the source system should be able to communicate to the replication interface on the SP of the destination system. To establish this, you must set up a switched network for remote replication. You can configure a replication interface with an IPv4 or IPv6 address.

    Replication is supported between like type IP addresses. For example, you can set up a replication connection between a replication interface configured with IPv4 address on the source system and a replication interface configured with IPv4 address on the remote system.

    Configuration for asynchronous replication

    When you create replication interfaces for asynchronous replication, use any of the available Ethernet ports to configure an IP address on each SP. These IP addresses are then used for asynchronous replication between the remote systems.

    Configuration for synchronous replication

    When you create replication interfaces for synchronous replication, you must use the Ethernet Port titled 'Sync Replication Management Port' to configure an IP address on each SP. This port is specifically used for management of the synchronous replication session, leveraging the same physical interface as the storage system's Unisphere management port, whereas the Fibre Channel (FC) ports are used for replicating data between systems. Before creating replication interfaces and connections for synchronous replication, ensure that you complete FC zoning and configuration.

    You can choose to configure Asynchronous, Synchronous, or both connection modes for block storage resources. If you choose to configure both connection modes, ensure that you configure an additional Ethernet Port IP interface on each SP of the source and destination systems.

    About replication connections

    Replication requires a configured connection between the source system and destination system. This connection is called a replication connection. Replication connections use replication interfaces on each SP from the local system and pairs them up with the replication interfaces on each SP in the remote system.

    For synchronous replication, a Fibre Channel (FC) connection is required for the data communication, and an IP interface for the management communication. Only one FC port can be configured and used for synchronous replication. Once an FC port is assigned, it will continue to act as the synchronous replication port and can never be changed. To determine the FC port used for synchronous replication, in the CLI console, run the command /remote/sys show -detail. Port information, similar to the following example, will appear in the output:

                        Synchronous FC ports = spb_fc4, spa_fc4
                      

    For more information, see the Unisphere CLI User Guide.

    Before creating a replication interface or connection for synchronous replication, it is recommended that you configure the FC zoning.
    Connection modes

    The following connection modes are supported:

    • Asynchronous – Supports asynchronous replication of file and block storage resources. It uses the Ethernet Port IP interfaces configured on Storage Processor A (SP A) and Storage Processor B (SP B) on the source and destination systems.
    • Synchronous – Supports synchronous replication of block storage resources only. It uses the Sync Replication Management Port with IP addresses configured on SP A and SP B on the source and destination systems.
    • Both – Supports synchronous replication of block storage resources using the Sync Replication Management Port and asynchronous replication of block storage resources. Ensure that you set up the additional Ethernet Port IP interfaces on SP A and SP B on the source and destination systems.

    About replication sessions

    A replication session leverages a configured replication connection and associated interfaces to establish an end-to-end path for replicating data between the source and destination storage resources. Once a replication session is established, the system creates two replication system snapshots on the source storage resource and two replication system snapshots on the destination storage resources. These snapshots are only used for asynchronous replication. The snapshots are used to keep point-in-time copies of data and maintain a common base between the source and destination storage resources. The snapshots are refreshed based on the replication mode:

    • For automatic synchronization, the snapshots are refreshed based on the Recovery Point Objective (RPO) you specified. You can adjust the RPO interval based on the amount of protection you want. You can specify an RPO from a minimum of 5 minutes up to a maximum of 1440 minutes (24 hours).
    • For manual synchronization, it is recommended that you sync the replication periodically.
    The snapshots consume protection space based on the snapshots feature. Although the internal snapshots may appear listed in the Snapshots tab of a storage resource (source or destination), you cannot manage or modify them. The snapshots also do not participate in automatic delete operations.

    The following steps provide a high level overview of how the replication session works:

    1. When the session and the internal snapshots are created, an initial synchronization occurs where all the data from the first snapshot of the source storage resource is replicated to the destination storage resource.
      For automatic synchronization, initial synchronization is enabled by default. For manual synchronization, there is no initial synchronization. An option is available to replicate existing snapshots on the storage resource as part of the initial synchronization. This option is only available when creating a replication session. In this case, replication transfers the snapshots in the order in which they were created on the source to the destination site. If you opt not to synchronize existing snapshots as part of the initial synchronization, the replication copies only the source data from the internal system snapshot.

      The system supports up to 16 concurrent initial synchronization operations per SP.

    2. Once data is replicated to the destination storage resource, the first snapshot on the destination storage resource is refreshed.

      The first snapshot on both systems then becomes your common base at that point in time. In event of a disaster, the common base is used to restore data from the last time it was synchronized.

    3. When the session is synchronized next, manually or automatically (based on the RPO), the second internal snapshot on the source system gets refreshed.
      If a replication session is created with auto synchronization, any snapshots created or marked for replication since the previous RPO are replicated upon RPO expiry to the destination site. In case of a session configured with manual synchronization, any snapshots created or marked for replication since the last synchronization occurred are replicated whenever the session is manually synchronized.
    4. Once the snapshot is refreshed, data is replicated to the destination storage resource.
      Only the differential data between the first and second snapshots gets copied over.
    5. The second snapshot on the destination storage resource is then refreshed.
    6. Once this occurs, at that point-in-time, the second snapshots on both the systems now become the common base.

    These steps occur each time a synchronization occurs. When a disaster occurs, replication uses the common base to go back to a good point in time copy of data. If failing over to the common base is not sufficient and replicated user snapshots exist, manually restore the destination resource to any of the replicated user snapshots.

    About replication support for thin storage resources with compression

    A thin destination LUN or Consistency Group (CG) of LUNs can be created with compression during replication session creation when the following criteria is met:

    • The source LUN or CG is thin.
    • The OE version is 4.2 or later.
    • The pool is all Flash.

    If the source LUN or CG meet this criteria, thin is selected by default for the destination LUN or CG. You can choose to unselect it to create a thick destination LUN or CG. If thin is not selected for the destination storage resource, compression cannot be selected. Compression can only be selected when thin can be and is selected for the destination storage resource.

    For file systems, replication matches the destination storage resource to the source. The destination is converted to match with the source after the synchronization operation (initial synchronization or manual synchronization). Therefore in this case, thin and compression cannot be selected.