NetVault Plug-in for SQL Server 11.2 - User Guide

Backing up replicated databases

Database replication is supported from SQL Server® 2005 onward. Plug‑in for SQL Server supports all three types of replication: Transactional Replication, Merge Replication, and Snapshot Replication.

Replicated databases and their associated system databases should be backed up regularly, including the following:

The replicated databases can be backed up in the same way as the non-replicated databases with the plug-in. If you perform regular Incremental Transaction Log backups, any replication-related changes should be captured in the Incremental Transaction Log backups. If you do not perform Incremental Transaction Log backups, a backup should be performed whenever a setting relevant to replication is changed.

For more information, see Strategies for Backing Up and Restoring Snapshot and Transactional Replication, Strategies for Backing Up and Restoring Merge Replication, and Common Actions Requiring an Updated Backup in the SQL Server Books Online.

Reviewing the compression features

Before configuring a backup, review the following information. This information applies to all versions of SQL Server® that the plug-in supports, although some of the features described are not supported by all versions of SQL Server.

If your version of SQL Server supports it, you can use the SQL Server Backup Compression feature. The plug-in also provides three levels of compression, low, medium, and high, which you can use whether or not your server supports SQL Server Backup Compression. You can also turn off compression. The compression options include:

To view the compression ratio achieved after a backup job has run, click Job Status in the Navigation pane, and select the completed job, and click View logs. In the log that appears, locate a message that starts with “Compression Ratio.” To view additional information, select the line, and click More info. You can then view the total number of bytes of uncompressed data that was backed up, the number of bytes that was sent to the backup media, and the number of administrative bytes included as a header, which is also included as part of the compressed data statistic but are extraneous to the actual compressed data. The compression ratio that is reported in the log is [[(total uncompressed data) – (total compressed data)] x 100%] / (total uncompressed data). The number of header bytes is usually an insignificant part of the backup size, except when small databases are backed up using multiple NetVault Backup streams.

If you use the SQL Server Backup Compression feature, or if you choose not to use compression, the VDI Streams when NetVault Backup Compression is not used option is enabled.

If you use one of the NetVault Backup Compression methods, the VDI Streams/Compression Threads and Number of Parallel NetVault Backup Streams options are enabled. The VDI threads perform compression. During a backup, each VDI thread or connection accepts data from SQL Server, compresses it, and then writes it to the output device using a NetVault Backup stream. Because higher compression levels are more processor-intensive, consider making effective use of the number of processors available on the system running the NetVault Backup Client, that is, the system running the SQL Server Instance that you are backing up. For example, you might specify a number of VDI threads that is slightly less than the number of processors on your system. However, the number of backup streams should not exceed the number of output devices. For this reason, you must specify the number of VDI streams (VDI Streams/Compression Threads) and the number of NetVault Backup streams (Number of Parallel NetVault Backup Streams) separately. When you access the SQL Server Backup Options tab, the number of VDI streams initially displayed is one less than the number of processors detected on the NetVault Backup Client. The value specified in this field determines the number of VDI devices to be used for storing the backup job. In the default setting, all backup job options have one VDI device set to the job, when Use Virtual Device Interface (VDI) is selected. You can add additional VDI devices; however, the minimum number of VDI devices is 1, and the maximum number cannot exceed 64.

IMPORTANT: If you use the SQL Server Backup Compression feature or no compression, the number of VDI devices selected must be equal to or less than the number of available media or tape drives. For example, if the number of VDI devices selected is 5, a minimum of five media or tape drives must exist. If you use NetVault Backup Compression, the number of parallel backup streams selected must be equal to or less than the number of available media or tape drives.

At the bottom of the VDI Backup Options section, the plug-in displays the number of processors on the NetVault Backup Client. If the processors use Hyper-Threading (HT), the number reflects it; a system with four physical processors and with HT enabled is identified as having eight processors. You can use this number when choosing the number of VDI streams if you use NetVault Backup Compression.

For optimal backup times when NetVault Backup Compression is used, base the entry for the Number of Parallel NetVault Backup Streams on the number of backup media. If your media uses striped disks, the fastest backup might occur with the number of backup streams set to the number of stripes. If the backup is writing to different hard drives that are not striped, with each NetVault Backup output device on a different hard drive, set the number of backup streams to the number of hard drives. If you are backing up to one or more VTLs on a single non-striped disk, use only one or two NetVault Backup streams. If you are backing up to a tape library with several tape drives, use no more backup streams, preferably fewer, than there are tape drives.

Whether or not you use NetVault Backup Compression, the number of VDI devices required for a restore is equal to the number that were used in the backup, and the number of backup streams required for the restore is also equal to the number that were used in the backup. If you use tape media and fewer tape drives are available for the restore than were used for the backup, the restore might fail. To avoid failure, Quest recommends that you use fewer NetVault Backup streams than you have tape drives available.

Performing Online VDI backups

An Online VDI backup using Plug‑in for SQL Server includes the steps outlined in the following topics.

Selecting data for an Online VDI backup

You must use sets — Backup Selection Set, Backup Options Set, Schedule Set, Target Set, and Advanced Options Set — to create a backup job.

Backup Selection Sets are essential for Incremental and Differential Backups. Create the Backup Selection Set during a Full Backup, and use it for Full, Incremental, and Differential Backups. The backup job reports an error if you do not use a Selection Set for the Incremental or Differential Backup. For more information, see the Quest NetVault Backup Administrator’s Guide.

TIP: To use an existing set, click Create Backup Job, and select the set from the Selections list.

IMPORTANT: If you are backing up data that was set up for use in a Virtual Client, select the SQL Server® Virtual Server, or the items contained within, for inclusion in the backup. If you are using multiple SQL Server Virtual Servers in the same cluster, select the correct SQL Server Virtual Server whose data you want to back up. The instance name displayed on this page is the SQL Server Virtual Server that was established as the Virtual Client during installation. Check that you run the process using the Virtual Client and not from one of the nodes. If you open or expand one of the nodes and drill down through the hierarchy, you might see a dimmed SQL Server Instance identified as local. While the system might use this instance in maintaining log information, do not run any processes at this level.

IMPORTANT: For the backup types Full File and Filegroup, Differential File and Filegroup, Partial Database, and Differential Partial Database, a read-only filegroup must be explicitly selected in the selection tree to include it in the backup. Though a green check mark appears against the read-only filegroups when you select the parent-level database, these filegroups are not backed up. When including a read-only filegroup in a backup, select the individual items instead of the parent-level database.