Frequently Asked Questions
Can I perform a synthetic full backup to one tape when the storage policy includes multiple clients?
No. The sequential-access nature of tape does not allow writing data onto its own source medium.
Client A, Client B, and Client C all write to the same storage policy. Each has completed a full backup and several incremental backups. A synthetic full backup of Client A must write to a new tape, since a source backup for Client A exists on the current active destination tape. The new tape becomes the active tape for all clients.
Immediately after the synthetic full backup of Client A, you run a synthetic full for Client B. Client B has not performed any backups since Client A's synthetic full, so it can write to the current active tape.
Next, you run an incremental backup of all clients. A new synthetic full backup of Client C requires another new tape, since its last incremental backup is on the current active tape with Client A and Client B's synthetic full backups.
Can I run an incremental backup job while a synthetic full backup is running?
File System Clients
Yes. You can manually run or schedule incremental backup jobs when a synthetic full backup is running.
Virtual Server Agents
For virtual server agents with Indexing Version 2 enabled, you can manually run or schedule incremental backup jobs while a synthetic full backup is running.
You do not need to schedule incremental backups around when synthetic full backups are performed. However, running a synthetic full backup while an incremental backup job is running is not supported. Also, running a full or differential backup job in parallel with a synthetic full backup job is not supported.
Does a synthetic full backup affect restore performance?
Commvault’s recovery process uses a single pass restore that only reads each object/block once as needed. Only the most recent object/block relative to the recovery point in time is read. On disk media, the object/block may be written anywhere free space is available. Restore performance is largely based on the random access speed of the disks. Frequent full backup or synthetic full backup jobs provide little benefit in recovery time.
However, job-based restores from tape media can benefit from full backups or synthetic full backups as these jobs actually consolidate data on the tape media. This consolidation can reduce load/seek time and improve read efficiency.
If I have to run a synthetic full backup on a certain day/time, how do I include the most recent incremental backup data?
You can do this by running incremental backups sequentially each day of the week, and then running a synthetic full backup after the final incremental job is run.
Suppose you want to run a synthetic full backup every Sunday at 6:00pm, but you want to ensure that your daily incremental backups are included in that synthetic full backup.
Perform the following steps:
- Configure an incremental backup schedule pattern on a backup set or subclient so that it runs at a specific time every day (for example, Monday through Sunday at 6:00pm). For more information, see Adding an Automatic Schedule for Incremental Backup.
- Configure a synthetic full backup schedule pattern on a backup set or subclient so that it runs at a specific time after the final incremental backup in Step 1, above, was run (for example, Sunday at 6:30pm). For more information, see Adding an Automatic Schedule for Synthetic Full Backup to a Backup Set or Subclient.
- To ensure that the synthetic full backup job does not conflict with the final incremental backup job in Step 1, make sure to that jobs are set to be queued. For more information, see Queuing Jobs.
Why should I perform a synthetic full backup?
First of all, you do not need to run a synthetic full to “true up” your backups. Any reconciliation of data not found/readable in protected storage is handled by subsequent backups. A synthetic full is required only when using subclient retention settings. In creating the list of objects to include in a synthetic full, subclient retention values are consulted.
For example, an archived object is deleted from the source, but subclient retention says to retain the object for a year. The deleted object will not be seen in subsequent incremental backup jobs so it’s not on their list of objects. The synthetic full looks at the previous backup’s list of objects and at the previous synthetic full or initial full backup. If subclient retention applies, deleted objects are added to the list of objects used for the next synthetic full. By this manner, a deleted or archived object is persisted in protected storage.
Besides the requirement for subclient retention, a synthetic full may be recommended when resources (for example, time, CPU, network bandwidth, etc.) required to do a full backup may not be available.
For example, a remote office is backing up servers over a network to corporate headquarters. A full backup of just one server takes nearly 10 hours due to available network bandwidth. An incremental backup only takes 15 minutes. In this scenario, performing a synthetic full backup would be prudent.
Last modified: 10/31/2019 8:03:58 PM