Tape Storage

General Remarks

Tape storage provides a versatile, long-term archival destination for DPX backups. By archiving backups to tape, organizations can benefit from a scalable, offline storage solution with built-in media management for handling retention, and recycling.

Tape libraries and drives can be directly attached to DPX servers via SAS or FC connections. They can also be accessed over a storage area network (SAN) or through NDMP-enabled NAS devices. This flexibility allows tape devices to be shared and utilized efficiently across multiple servers.

Tape drives and libraries are configured globally under Media Management in the DPX interface. Backup jobs can then be created to archive data to tape based on specified policies and schedules. DPX handles all media rotation, recycling, and retention according to predefined rules. Critical capabilities like tape indexing and caching allow for fast searches and restores.

Tape storage can be deployed as:

• Attached directly to a DPX server

• Shared over a SAN

• Connected to a NAS device that is NDMP-enabled

Tape Multistreaming and Multiplexing

Catalogic DPX supports tape multistreaming and tape multiplexing – two methods to optimize performance by improving data transfer and utilizing tape drive storage resources more efficiently.

Note. Multistreaming and multiplexing are designed to optimize backups to physical tape libraries. While the logic also applies to virtual tape libraries (DiskDirectory), the benefits may not compensate for the costs, such as higher demand for resources.

Tape Multistreaming

This process splits a single backup job into multiple concurrent streams, each writing to a separate tape drive. It speeds up the backup process for large datasets but requires more resources.

Multistreaming is enabled at the job definition level and is supported for file backup jobs.

Web interface

To enable streaming in file backup:

1. Create a file backup job definition with at least two different nodes as sources.

2. Scroll down and select Advanced Options > Destination Options. You can separately set destination options for base, incremental and differential backup.

1. Select a value from the Performance Options dropdown. This option helps control the number of concurrent tasks within a backup job.

Option	Description
Standard	Backs up UNIX nodes and Windows disks concurrently. Use this option when memory on the master or device(s) is limited or when you will not derive benefit from any of the other choices.
Split Job by Partitions	Backs up UNIX partitions (mount points) and Windows disks concurrently. Use this option when backing up multiple partitions and there are no device or memory limitations.
Split Job by Files	Use this option when backing up single partitions (disks or volumes) that contain many files. This selection will create a task for every available tape drive in this job. Note. This option is not supported for Linux OES.
Split Large Files/Partitions	Splits files that exceed the Split Size threshold. Each part of the split file is backed up concurrently. Use this with large raw partition backups or when you are backing up very large files and the number of files or partitions is less than the number of devices.

Desktop interface

To enable multistreaming in file backup:

1. Create a file backup job definition with at least two different nodes as sources.

2. In Other Tasks section, select Set Destination Options.

1. In the Set Job Destination Options dialog, select an option as per description below. You can separately set destination options for base, incremental and differential backup in their relative tabs.

Option	Description
Standard	Backs up UNIX nodes and Windows disks concurrently. Use this option when memory on the master or device(s) is limited or when you will not derive benefit from any of the other choices.
Split Job by Partitions	Backs up UNIX partitions (mount points) and Windows disks concurrently. Use this option when backing up multiple partitions and there are no device or memory limitations.
Split Job by Files	Use this option when backing up single partitions (disks or volumes) that contain many files. This selection will create a task for every available tape drive in this job. Note. This option is not supported for Linux OES.
Split Large Files/Partitions	Splits files that exceed the Split Size threshold. Each part of the split file is backed up concurrently. Use this with large raw partition backups or when you are backing up very large files and the number of files or partitions is less than the number of devices.

Tape Multiplexing

In this process, multiple backup streams from different sources are written concurrently to a single tape drive or DiskDirectory. Data blocks from various nodes are interleaved within a single tape stream, optimizing backup time. This approach also reduces tape wear by allowing the tape to run at a stable speed without frequent stops to buffer and write data.

Multiplexing is enabled at the job definition level and is supported for file backup jobs.

Note. Increasing concurrency to values >1 during backups slows restore performance and requires more memory on the master server during restore.

Web interface

To enable multiplexing in file backup:

1. Create a file backup job definition with at least two different nodes as sources.

2. Scroll down and select Advanced Options > Destination Options. You can separately set destination options for base, incremental and differential backup.

1. Change the value in the Concurrency per Device field to different from 1. This field sets the maximum number of file backup tasks to a single device.

Desktop interface

To enable multiplexing in file backup:

1. Create a file backup job definition with at least two different nodes as sources.

2. In Other Tasks section, select Set Destination Options.

1. In the Set Job Destination Options dialog, change the value in the Concurrency per Device field to different from 1. This field sets the maximum number of file backup tasks to a single device. You can separately set destination options for base, incremental and differential backup in their relative tabs.