Backing up the integrated file system

Memory management

The SAV command provides the Asynchronous bring (ASYNCBRING) parameter which allows you to modify the memory management method that is used to access the data. Bringing objects into memory asynchronously may improve the performance of a save operation.

Using concurrent backup operations

Reduce your backup windows by using multiple concurrent backups. To implement this approach you will need to determine some way to group your integrated file system data. Then you need separate SAV commands to concurrently save each of the subsets. You will need to consider the potential resource contention that can occur on the hardware resources being used. For example, performing concurrent backups on groups of data that are stored on the same set of disk units might cause contention on those disk units. You might decide to use multiple tape drives or a tape library system with multiple drives to run multiple concurrent SAV commands.

For more information about concurrent backups, see Saving to multiple devices to reduce your save window.

Saving to save files (SAVF) then saving the SAVFs to tape with SAVSAVFDTA

Some customers have found that they can reduce their backup window by first backing up their data to a save file (SAVF) rather than saving directly to tape. Significant performance improvements were made to backups to save files. Of course if you back up to a save file, you need to have adequate disk space available for the save file. The Performance Capabilities Reference can help you evaluate this approach on your system. You also will need to back up your save files to tape by using the Save Save File Data (SAVSAVFDTA) command. However, the SAVSAVFDTA command does not need to be completed during your backup window.

Reducing or eliminating auditing during backup or recovery operations

Measurements show that performing security auditing during save or restore operations (*SAVRST auditing) can decrease performance. Auditing provides valuable information about the actions being performed on your system and who is performing those actions. However, you need to balance the value of that information against the time you have available to perform a backup or recovery. This is especially true if you need to recover all or many objects in the integrated file system.

Reducing the number of objects scanned during the backup

Specifying the SCAN parameter during the backup might have significant performance impacts if scanning is enabled on the system. Scanning objects might be a valuable part of your system security, but you need to consider the amount of time scanning will add to your backup window.

Using BRMS online backup of Lotus servers

Backup, Recovery, and Media Services (BRMS) supports online backups of Lotus® server databases (such as Lotus Domino® 8 for i and Quickplace). An online backup is a backup that you do while your Lotus server databases are in use; there are no save-while-active synchronization points. You can direct your online backups to a tape device, media library, save file, or a Tivoli® Storage Manager server. BRMS can also create control groups that make it easy to use concurrent backups. Performing an online backup does not improve the performance of your backup. However, since your applications remain active, the duration of the backup is less important.

For more information about BRMS online backups, see Backup, Recovery, and Media Services.

Using save-while-active

The SAV command provides the SAVACT, SAVACTMSGQ, and SAVACTOPT parameters to support saving objects while active.

For more information, see Saving your system while it is active.

Using the CHGPERIOD parameter to only save changed objects

The SAV command provides a CHGPERIOD parameter that can be used to find and save only objects that have changed. In some cases, this can be an effective way to reduce the amount of data you need to back up. However, the system still needs to look at each object to determine which objects have changed. If you have many files it might still take a long time to determine which objects have changed.

Structuring your directories to easily back up new files, omit data, or group your data

It might be beneficial to consider your backup strategy when you structure and name your directories. You might be able to group and name your files in some way that will make it easier to include or omit groups of directories or objects from your backups. You might want to group the directories such that you can back up all of the directories and files for an application, a user, or specified time period.

For example, if you are creating many files each day or each week, it might be useful to create a directory to contain the new files. Consider using a naming convention for the directories such that you can back up only the directory that contains the new objects or omit older directories.

/2008
/2008/01
/2008/01/01
/2008/01/02
/2008/01/03
/2008/01/04
/2008/02

Omitting objects from the backup

The SAV command provides the OBJ parameter that specifies the objects to be included and omitted from the backup. The OBJ parameter lets you specify a list of 300 values to be included or omitted from the SAV command. The values can be either specific directories or objects or generic values that provide wildcard support for the objects to be included or omitted.

Pattern parameter

The SAV command provides the PATTERN parameter which lets you specify a list of 300 values which are used to group the save by either including or omitting objects which qualify for the save based on the OBJ parameter. The values can be either specific object names or generic values that provide wildcard support for the objects to be included or omitted.

Journal changes and save journal receivers

You can journal changes to directories, stream files, and symbolic links. If you set up journaling on your integrated file system data, you might need to change your save strategy. Your new strategy should be to back up the objects less frequently and instead back up the journal receivers that contain the changes you've made to the objects. This could reduce the amount of data you need to back up. However, you will need to understand and consider the implications to your recovery procedures.

Implement Hierarchical Storage Management (HSM)

If you have historical integrated file system data that is infrequently needed you might benefit from using Hierarchical Storage Management. Hierarchical Storage Management (HSM) automatically and transparently manages customer data across a storage hierarchy. The storage hierarchy can consist of high performance disk, compressed disk, and tape libraries.

When and how often data is accessed on your system depends on the type of data. A set of data that is currently being used might be accessed many times a day (hot data), or it might have become historical data which is accessed less frequently (cold data).

Through the Backup, Recovery, and Media Services (BRMS) user-defined policies, HSM can migrate or archive and dynamically retrieve infrequently used data or historical data up or down a hierarchy of storage devices.