Replication

The DXi9000 Series provides data replication capabilities that you can use as an integral part of a disaster recovery plan. Replication allows you to configure the DXi9000 Series to create a copy of your data on any other DXi system on a compatible software version.

Note: Quantum strongly advises creating a dedicated share for each backup application. Not only is this easiest for storage management, but some backup applications do not release files immediately, so there is a potential for conflicts, particularly with replicated shares.

In the event of adisaster or ransomware attack, you can quickly recover the replicated data on the remote system, allowing your business to resume normal operations. Once the original system is available again, you can restore all data back to its original location.

Note: If you have more than one IP address configured on you DXi network page, it is highly recommended to implement static routes to the replication partner (both from source to target and from target to source). See Understanding Interface Routing for more information.

Additional Information

During data replication, data is sent from one system (the source) to another system, usually in another location (the target). For example, you might replicate data from a branch office to a central office.

Replication works only with deduplicated data, and data is compressed before it is replicated, or when a tape cartridge is unmounted (VTL partitions) replicated. Because of this, the amount of data transmitted between systems during replication is greatly reduced compared to the original amount of data stored. In addition, a data block is transmitted only if the target does not already have a copy of the block. Finally, data can optionally be encrypted before it is transmitted.

DXi supports two types of replication:

Namespace replication works by constantly sending deduplicated data to the target system, and then either by schedule or manually, creating a snapshot on the target system. this gives you a point-in-time snapshot of the stored data that that remains until the snapshot is deleted. The number of snapshots that are maintained is controlled by the receiving system.

If the sending DXi is lost or data is corrupted, a "restore" operation can be run from the receiving DXi to re-create the data exactly as it was at the time the snapshot was sent.

After you recover a share, it is re-created on the target system and is available for use. After the source system becomes available again, you can perform a failback operation to restore the share to its original location.

Caution: A saved snapshot is necessary to recover your data at a later time. For this reason, it is not enough to simply enable replication for a share or partition. You must also configure a replication schedule (recommended) or perform manual replication on a regular basis to send snapshots of the share or partition to the target system.

Like Namespace replication, Directory/File or Cartridge Based Replication constantly sends data from a share on the source to a share on another DXi. But it is different in a number of ways:

  • Data is mirrored between the source and target systems and both have all the same files available at the same time (no snapshots for a point in time).
  • A unique Sync ID is used to associate the replicated share or partition on the source system with the share or partition that will receive the replicated data on the target system.

  • You do not need to schedule or manually perform share or partition through the remote management console. Instead, replication is triggered when a file is closed or a period of time after it is modified (NAS shares).

  • After files or cartridges are replicated using Directory/File or Cartridge Based Replication, they are automatically recovered on the target system and are immediately available for use. There is no need to first recover the share or partition to access its data.
  • Deletions are automatically propagated from the source system to the target system in order to free up space on the target. In addition, you can manually initiate a synchronization from the source system at any time. Synchronizing ensures that the contents of the source share or partition are exactly the same as the target share or partition.

Caution: Cartridge barcodes for all source systems that are replicating to the same target must be unique.

Additional Information

  • For CIFS/SMB shares, a file is replicated immediately after it is closed. For NFS shares, a file is replicated after it has been idle (not accessed) for several minutes.

  • To optimize the replication process, deduplicated data is continuously sent in the background from the source system to the target system. Any remaining data is sent (along with metadata) at the time a file is closed.

  • If you delete a share configured for Directory/File-based Replication on the source DXi, the share is not automatically deleted on the target. If you do not want to retain the share on the target DXi, you can manually delete it.

  • Snapshots for a point in time can be created and protected on the target replication system using Secure Snapshots. But deleting these snapshots is not allowed and they must expire based on retention days.

The DXi9000 Series can duplicate the data on an LSU (logical storage unit) to another location using the OST optimized duplication (OST replication) feature supported by Veritas NetBackup and Backup Exec. If you are using Veritas NetBackup 7.1 or higher, you can configure an LSU for Automatic Image Replication.

For information about configuring and using OST optimized duplication and Automatic Image Replication, see Replicating OST Data.

The DXi9000 Series can send replicated data to multiple target systems. First, configure up to two replication targets. Then, for each share or partition, select which targets where you will replicate data, to one, both, or neither. See Adding a Replication Target and Enabling Replication For a Share or Partition.

Below are three possible scenarios in which multiple replication targets are configured.

All Shares or Partitions Replicate to Multiple Targets

For enhanced disaster recovery, all shares or partitions on the DXi9000 Series are configured to replicate to multiple targets. In this scenario, even if one target system is lost, the other target system still retains a full copy of all data on the source DXi9000 Series.

Some Shares or Partitions Replicate to Multiple Targets

For increased flexibility, each share or partition on the DXi9000 Series is configured to replicate only to the targets where needed. In this scenario, share or partition fall into three categories:

  • Some shares or partitions replicate to Target 1. (They may also replicate to Target 2.)
  • Some shares or partitions replicate to Target 2. (They may also replicate to Target 1.)
  • Some shares or partitions are not replicated.

The Replication page allows you to set up replication for NAS, VTL, or Application Specific shares, and to replicate and recover data.

To access the Replication page, click the Replication menu.

The Replication page contains the following tabs:

Task Overview

Use the DXi9000 Series data replication features to perform the following tasks:

To replicate all data on a NAS share or VTL partition from a source system to a target system:

  1. On the target system, add the source system to the list of allowed replication sources. See Adding a Replication Source.
  2. On the source system, specify the target system that will receive replicated data. See Adding a Replication Target.
  3. On the source system, create a new share or partition with deduplication enabled. See Adding a NAS Share or Adding a VTL Partition.
  4. Enable replication for the new share or partition. See Enabling Replication For a Share or Partition.

  5. Before writing any data to the new share or partition, replicate the new share or partition. See Replicating a Share or Partition.

    Quantum recommends that you always replicate a new share or partition immediately after creating it. This establishes the initial data structure on the target system and greatly increases the speed of the first replication.

  6. Choose one of the following methods to regularly replicate the new share or partition:

After replication, a snapshot is sent to the target system. You can use the snapshot to recover the replicated share or partition on the target system or restore the share or partition to its original location on the source system. See Task Overview: Recovering a Namespace Replicated Share or Partition and Task Overview: Performing a share or partition Failback.

Important Information: Setting up and Performing Replication

  • If you are setting up "namespace" type replication and do not configure a replication schedule or manually replicate the share or partition on a regular basis, your data is not protected. While the DXi9000 Series continually sends data to the target system in the background to optimize the replication process, a snapshot that preserves the complete structure of your data is sent to the target system only when a scheduled or manual replication job occurs. If you do not regularly and frequently replicate your data as described in Step 6 above, it cannot be restored at a later time.
  • If a replication is not scheduled or run manually for namespace-type replication, data is still continuously sent to the target once replication is configured on the source DXi9000 Series. This can potentially fill up the target DXi9000 Series.

Note: To limit the amount of network bandwidth used for replication, enable a constant throttle or schedule replication throttling. See Enabling System Throttling and Scheduling Replication Throttling.

To automatically replicate files or cartridges from a source system and immediately recover them for use on a target system:

  1. Perform the following steps on the target system:

    1. Add the source system to the list of allowed replication sources. See Adding a Replication Source.
    2. Create a new share or partition with deduplication enabled. See Adding a NAS Share or Adding a VTL Partition.
    3. Enable Directory/File or Cartridge Based Replication for the new share or partition and specify the Sync ID. See Configuring Directory/File or Cartridge Based Replication.

  2. Perform the following steps on the source system:

    1. Specify the target system that will receive replicated data. See Adding a Replication Target.
    2. Create a new share or partition with deduplication enabled. See Adding a NAS Share or Adding a VTL Partition.
    3. Enable Directory/File or Cartridge Based Replication for the new share or partition and specify the Sync ID. See Configuring Directory/File or Cartridge Based Replication.

A file is automatically replicated when it is closed or a period of time after it is modified. A cartridge is automatically replicated after it is unmounted. After replication, the replicated files or cartridges are immediately available on the target system. There is no need to recover the share or partition to make the files or cartridges available.

Caution: Cartridge barcodes for all source systems that are replicating to the same target must be unique.

Note: To limit the amount of network bandwidth used for replication, enable a constant throttle, or schedule replication throttling. See Enabling System Throttling and Scheduling Replication Throttling.

If a NAS share or VTL partition is lost or damaged on the source system, you can recover it on the target system using a received snapshot. When you recover a share or partition, it is recreated on the target system just as it was at the point in time when the snapshot was saved.

To recover a replicated share or partition:

  1. On the target system, select a received snapshot and perform a recover operation. See Recovering a Share or Recovering a Partition.

    The share or partition is now available on the target system.

  2. (Optional) Map your backup application to the recovered share or partition on the target system to continue making backups.

  3. When the source system is operating correctly again, choose one of the following actions:

    • Failback the share or partition to restore it on the original source system. Then, if necessary, map your backup application to the restored share or partition on the original source system. See Task Overview: Performing a share or partition Failback.
    • Continue to make backups to the recovered share or partition on the original target system. In addition, set up replication to replicate the share or partition back to the original source system. In this scenario, the original source system is now the target, and the original target system is now the source. See Task Overview: Setting Up and Performing Namespace Replication.

Note: When you failback a share or partition, the snapshot from the point in time of the original replication is replicated back to the original replication source. This snapshot then needs to be recovered. See Recovering a Share or Recovering a Partition.

If a NAS share or VTL partition is lost or damaged on the source system, you can failback the share or partition to the source system using a received snapshot on the target system. When you failback a share or partition, it is restored on the source system just as it was at the point in time when the snapshot was saved.

To perform a share or partition failback:

  1. On the original source system, add the original target system to the list of allowed replication sources. See Adding a Replication Source.

    Note: For the purposes of failback, the original source system is now acting as a target because it will receive the snapshot from the original target system.

  2. On the original target system, select a received snapshot and perform a failback operation pointing to the original source system, and then recover the share or partition on the original source system. See Performing a Failback For a Share or Performing a Failback For a Partition.

    The share or partition is now available on the original source system.

  3. (Optional) If necessary, map your backup application to the restored share or partition on the original source system.