Synthetic Full Backups

Overview

As the name suggests, a synthetic full backup is a synthesized backup. It is created from the most recent full backup (i.e., standard or synthetic) and subsequent incremental and/or differential backups. It, therefore, cannot be the first backup job of a client computer or subclient. The resulting synthetic full backup is identical to what would have been created had the last backup for the subclient been a full backup.

Unlike full, incremental, and differential backups, a synthetic full backup does not actually transfer data from a client computer to the backup media. In this sense it is not really a backup at all, rather a backup consolidator. Therefore, since synthetic full backups do not back up data from the client computer, they  impose no load on the client computer.

Because synthetic full backups consolidate backup data into one archive file, they provide a means of enhancing restore performance, particularly when a given full backup cycle contains many incremental backups.

Synthetic full backups are treated by the system as standard full backups, but cannot be the next backup job after a standard full backup. As a result, they delineate full backup cycles in the same manner as standard backups. This is an important point to remember, since retention periods are defined in part by the number of full backup cycles. Using synthetic full backups in an unguided manner may cause the unintentional expiration of data.

Synthetic full backups are media-based; they read backup data from one media and simultaneously write (restore) the result to the new active media within the same media group. As a result, synthetic full backups require that at least two media drives for the same storage policy be available at the time the job is started. Synthetic full backups cannot be performed on subclients where the storage policy is associated with a stand-alone drive.

How Synthetic Full Backups Work

You can start a synthetic full backup at the subclient, backup set or iDataAgent level. When started for a backup set or iDataAgent, the operation acts on each constituent subclient individually. In other words, if you start a synthetic full for a backup set that contains two subclients, the system automatically conducts two synthetic full backups, one for each subclient. Therefore, if you choose to use a new media for synthetic full backups, then you may need to have several spare media available.

A synthetic full backup extracts the index data of each participating subclient. Using this index data and the previously backed up user data images, it builds new full backup images, one for each subclient. The new backup images consolidate the index and user data stored in the related incremental, differential, and previous full backups. As the synthetic full backup for each subclient proceeds, the system writes an archive file to the storage policy from which the backup data originated (i.e., one archive file per subclient). Since each archive file represents one synthetic full backup for each subclient, a synthetic full backup of a backup set containing three subclients would initiate three operations, each resulting in an archive file.

The following illustration demonstrates the operation of a synthetic full backup that was initiated at the backup set level. For simplicity, assume:

• The backup set contains two subclients; default and subclient1
• Each subclient has the same backup schedule
• Index and user data of the backups are directed to the same storage policy. (This is the default behavior of standard backups.)
• Backups for each subclient are directed to separate and dedicated storage policies (i.e., groups that service no other subclients.)

From this figure, we can see that from a single request on 4/7, two synthetic full backups were created: one for each subclient. When we requested a synthetic full backup for the backup set, the system identified all the backups that were created for each subclient back to the most recent full (F), which in this case occurred on 4/1. Using the index data from these backups, the system created new consolidated index and user data images and wrote them to new archive files (SF) on the same storage policies.

In this example, we assumed that the standard backup was written to a dedicated storage policy and that the user data and corresponding index were both backed up to the same storage policy. Note that these conditions do not affect the overall operation of a synthetic full backup. In either case, (i.e., non-dedicated storage policies or divided index/user data) the system locates the applicable index and user data and creates a new archive file.

An important point to remember is that a synthetic full backup consolidates data; it does not back up data from the client computer. Consequently, the data in a synthetic full backup always reflects the data of the previous backup, regardless of when the synthetic full backup was requested. In this example, the synthetic full backup was started on 4/7; however, the data it contains is from the 4/1 through 4/6 backups. To secure data for 4/7, we would need to run a conventional (i.e., full, incremental, or differential) backup on 4/7.

Also, keep in mind that synthetic full reads the data from the copy whose restore copy precedence is set to 1. And like a real backup, synthetic full is always created in the primary copy. i.e., its destination is always a primary copy.

Running Synthetic Full Backups

The following figures show two examples of using synthetic full backups in standard backup schedules. The figures show a series of full backup cycles where each cycle is a week in length. The longer lines indicate full backups while the shorter lines indicate incrementals. Note that both sets of backups begin with the required standard full backup.

In the first figure, both synthetic and standard full backups are used to define the full backup cycles. Synthetic full backups are used more often because they do not affect the operation of the client computer. In the second figure, synthetic full backups are used exclusively. To ensure that client data is properly secured, each synthetic full backup is followed by an incremental backup on the same day. If you decide to incorporate synthetic full backups into your standard backup schedules, the approach you take is entirely dependent upon your needs.

If a tape library is being used for synthetic full backups with multiple streams where the number of streams is more than the number of drives on the tape, then its recommended to have only one synthetic full job running to the tape drive.

1. From the CommCell Browser, expand Client Computers by double-clicking Client Computers | File System iDataAgent | defaultBackupSet. The default and other subclients (if available) are displayed on the right-hand windowpane.
2. Right-click the default subclient, and click Backup.
3. From the Backup Options dialog box, select Run Immediately.
4. Select Synthetic Full backup.
5. Click OK. You can track the progress of the backup job from the Job Controller window. When the backup has completed, the Job Controller displays Completed.

After running a backup, you may want to verify the backup data. You can do this by viewing the Backup History. For more information, see Backup Job History.

Running an Incremental Backup Before or After a Synthetic Full Backup

An incremental backup can be run either before or after a synthetic full backup. Running an incremental backup immediately before the synthetic full ensures that any new or recently changed data is included in the synthetic full. Running an incremental backup immediately after the synthetic full ensures that any new or recently changed data since the backup that occurred prior to the synthetic full, but was not included in the synthetic full, is backed up by the incremental.

Verifying Synthetic Full Backups

When verifying synthetic full backups, disparities between actual files on the client computer and the Index are collected.  Internally, a flag is set when the synthetic full backup completes successfully. This flag adds functionality to the next incremental/differential backup to detect any items that the previous synthetic full backup did not include, and include any such items in that incremental/differential backup. The pending flag is cleared when the incremental/differential backup completes successfully, or when a conventional full backup completes, whichever occurs first.

1. From the CommCell Browser, select one of the following:
   • To back up a subclient, right-click the subclient and click Backup.
   • To back up a backup set or instance, right-click the backup set or instance, click All Tasks, and click Backup All Subclients.
   • To back up the default backup set, right-click the agent or instance node, click All Tasks, and click Backup Default Backup Set.
2. If you chose a level higher than subclient (i.e., backup set, etc.), you are prompted to confirm that you want to back up all the subclients below that level/node. Click Yes.
3. From the Backup Options dialog box, select Synthetic Full.  Optionally, select Run Incremental Backup and specify if it should be run before or after the synthetic full backup.
4. Click Advanced.
5. From the Advanced Backup Options dialog box, select Verify Synthetic Full. Click OK to continue.
6. Click OK to continue.

Errors encountered in synthetic full backups can be ignored, even is backup media is partially corrupted.  This is useful when running full backups is not possible and when synthetic full backups are continuously run. Errors such as inability to read data (bad tape, files missing on disk media, etc.) will be ignored when this option is enabled.

The next conventional backup must have the Verify Synthetic Full enabled  to re-backup the items that could not be copied by the synthetic full backup. See Verifying Synthetic Full Backup described above for step-by-step instructions.

1. From the CommCell Browser, select one of the following:
   • To back up a subclient, right-click the subclient and click Backup.
   • To back up a backup set or instance, right-click the backup set or instance, click All Tasks, and click Backup All Subclients.
   • To back up the default backup set, right-click the agent or instance node, click All Tasks, and click Backup Default Backup Set.
2. If you chose a level higher than subclient (i.e., backup set, etc.), you are prompted to confirm that you want to back up all the subclients below that level/node. Click Yes.
3. From the Backup Options dialog box, select Synthetic Full.
4. Click Advanced.
5. From the Advanced Backup Options dialog box, select Verify Synthetic Full and then select Ignore Error in Synthetic Full option.

   Ignore Error in Synthetic Full option can only be selected when Verify Synthetic Full is enabled.

6. Click OK .

Best Practices

Optimizing synthetic full backups in Deduplication appliances

Synthetic full backup operations back up all the data which are recently added or modified and writes a new baseline of items as a new job. This allows prior job to be aged, and then the new incremental backups are performed with this baseline as the reference.

Performance of a deduplication appliance such as Data Domain Array is bound by stream or thread performance. The synthetic full operation requires a concurrent read/write operation, which can affect the performance of the device.

It is recommended that you take the following steps to optimize the performance of the device:

• Stagger your backup schedule in such a way that the minimum number of active streams are targeted to the device.
• Alternately, if the synthetic full operation takes too long to complete, you can perform a Full backup followed by daily incremental backup jobs.
• In cases where longer backup cycles are performed in the device, the backup performance may degrade over time. This may be not be due to the concurrent read/write synthetic full jobs, instead it may be due to the large volume of deduplicated data on the deduplication appliance.