Active Full Backup
- The first run of the Backup Job will generate an Active Full Backup, getting all the data from the Production Storage.
- During the following days, the Job will generate incremental backups as configured in the Job.
- On Saturday, as scheduled, a new Active Full Backup is generated, obtaining again all the data from the Production Storage.
Synthetic Full Backup
- Because most of the data is obtained from existing backups in the repository, the use of network resources (SAN or LAN) is reduced. Remember that an Active Full Backup obtains all the VM data from the Datastore using the SAN or LAN, depending on the transport mode used.
- Similarly, because you don’t have to get data from the Production Storage, the impact over this Storage is reduced during backup operations, because most of the data will not be obtained from this Storage.
- On the other hand, a Synthetic Full Backup creates a greater impact on the Backup Repository, so, if the performance (IOPS / Latency) of the storage device used to store backups is limited (which happens very often), the creation of a Synthetic Full Backup could take a considerable amount of time ( See post about the Impact of backup methods in performance ) and also affect the performance of other Backup Jobs that are running at the same time as the Synthetic Full.
- Also the creation of Synthetic Full Backups using Deduplication appliances as a Repository is not recommended, because these kind of appliances are not optimized for read operations, and a Synthetic Full Backup requires to read the previous backup files to “synthesize” a new full backup.
- In a deduplication appliance, in order to read a data block, this must be re-hydrated, this means the appliance must undo the deduplication process applied to this data block.
- This generates a greater latency on read operations, and as a result, the required time to create a Synthetic Full Backup rises dramatically.
- The exception to the previous point is if you use a Deduplication Appliance with native integration with Veeam:
- EMC Data Domain with DDBoost
- HPE StoreOnce with Catalyst
- The day that Synthetic Full Backup is scheduled (Saturday), Veeam will run the Backup Job on a regular basis to generate the incremental backup scheduled to be created this same day.
- During the execution of the Job, Veeam will create this incremental backup in the regular way, this mean it will obtain the data from the Production Storage (Datastore), creating a new incremental backup in the existing backup chain (VIB).
- This incremental backup allows Veeam to make sure the Synthetic Full backup includes the latest changes in the VM that we are backing up.
- After the Job session ends, Veeam will build a new Synthetic Full Backup using the backup files that are already available in the repository, also adding the data obtained as part of the incremental backup indicated above.
- Additionally, in this step the incremental backup file (VIB) is eliminated, and instead only the Synthetic Full Backup (VBK) remains, which creates a new backup chain.
- The new incremental backups created later by the Backup Job are associated with this new Synthetic Full Backup in the same backup chain, until a new full backup (Active Full or Synthetic Full) is created.
Veeam treats Synthetic Full Backup in the same way that an Active Full Backup. Thus, when we ask for a Synthetic Full Backup on a Job, this will reset the backup chain, which means that all subsequent incremental backups are associated with this new Full Backup created “synthetically”. The previous backup chain (Full Backup + Incremental Backups) remains in the Repository until it is automatically deleted according to the retention policy configured in the Backup Job.
- The backup repository performance is constrained, so that the additional load generated by creating a Synthetic Full Backup creates a negative impact on the repository and on the time required to complete backup operations.
- Backup repository corresponds to a deduplication appliance , which doesn’t correspond to those with native integration with Veeam (EMC Data Domain with DDBost, with StoreOnce Catalyst HPE and ExaGrid)
- The performance of Production Storage is an important concern, so you should minimize the impact of running backups in order to avoid problems over the company business services.
- The available bandwidth in the SAN or LAN (Depending on the transport mode used) is limited, which requires reducing the amount of data transmitted during backup operations.
- This point is also important when you need to head a backup to a remote location where bandwidth is limited.