SEP sesam offers different types of backups, called backup levels, which allow you to define the level of data that is copied from the source to the destination. There are four different backup levels in SEP sesam: full, differential, incremental, and copy. The backup level is specified when creating a backup event in the Main Selection -> Scheduling -> Schedules -> New Backup Event.
Incremental and differential backups are typically used to reduce storage space as they only save the data created or changed after the last FULL backup (DIFF backup) or after the last backup (INCR backup) – whether FULL, DIFF or INCR – of the same task. A saveset created as FULL is the base saveset for all subsequent DIFF or INCR savesets, therefore you have to perform FULL backups regularly to ensure that you can always revert to the previous versions.
For example, a INCR backup that was taken after the third INCR after FULL, requires the FULL, the first, the second, and the third INCR to allow a full restore. If a saveset is missing from the backup chain, you cannot restore your data to a specific point in time.
Regardless of the backup level specified in the backup event, a backup may run as FULL (or DIFF) if one of the following conditions – explicit (option Enforce FULL) or implicit (ensure a valid backup chain) – is triggered.
- By enabling the Enforce FULL option when creating INCR and DIFF backup events, you can ensure that a FULL backup is run within the specified interval; see section Enabling the Enforce FULL option in the GUI.
- SEP sesam has a built-in mechanism to ensure that your chain of backup savesets always has the last full backup available. It automatically maintains control over dependencies between savesets and provides dependency-based automatic retention. For details, see the section Conditions for performing a FULL backup instead of a DIFF/INCR.
|A backup chain (the full backup and all subsequent incremental backups) is treated by retention management as if it were a single backup. Note, however, that there is currently no limit to backup chains in SEP sesam. Since long backup chains are inefficient and more vulnerable if something goes wrong because the backups in the chain depend on each other, and because they can simply become too long to restore (no more storage available, takes too long), SEP AG strongly recommends setting up regular FULL backups to avoid such problems, and setting the limit to up to 100 incremental backups.|
Enabling the Enforce FULL option in the GUI
For details on creating a backup event, see Creating a Backup Event. This section only provides information related to the option Enforce FULL.
- From Main Selection -> Scheduling -> Schedules, right-click the schedule for which you want to create a new event, then click New Backup Event. If you have already created a backup event, select the relevant schedule and then double-click the backup event to open its properties. The Parameter dialog opens automatically.
- Under Object, select the task for which you are creating this backup event from the drop-down list.
- Under Parameter, set the Backup level: select DIFF or INCR. You should consider the advantages of each backup level (regarding time and resources required to store and restore your data) when planning your backup strategy.
- Once your backup level is set to DIFF or INCR, you can enable the Enforce FULL option. This option ensures that a FULL backup is performed within the specified interval. For example, setting the number to 7 will enforce a full backup every seven days instead of an incremental or differential backup if the scheduled full backup fails. This ensures that a full backup is always performed within the specified time interval, in our example, once a week. You should consider the advantages and disadvantages of enforced full backup frequency, as it is slower than other backup levels and requires the most storage space.
If the value is set to 0, the system checks whether the last FULL backup was successful or run with warning each time a backup is performed. If the last FULL backup was terminated or run with error, the DIFF/INCR backup to be started is automatically executed as FULL backup.
|If no initial FULL backup exists, differential (DIFF) or incremental (INCR) backups are automatically performed as FULL backups.|
Conditions for performing a FULL backup instead of a DIFF/INCR
In addition to the option Enforce FULL, which can be activated manually in the GUI (see section above), there are also specific circumstances that will automatically trigger a full backup. SEP sesam performs a FULL backup instead of a DIFF/INCR when the following happens:
- copy_reformat_lis failed
- For example, if *.lis files cannot be transferred from RDS to SEP sesam Server. *.lis files contain information about the backup's unique files; once the backup is finished, these files have to be copied from RDS to the SEP sesam
SESAM_VAR/lisdirectory as this data is needed for a selective restore.
- The previous backup chain (FULL–DIFF–INCR) is not complete
- For example, if the EOL of a saveset is gone or a saveset is missing, SEP sesam will enforce a FULL backup. For more details on retention behavior and the different EOL parameters, see Automatic Retention (EOL) Management.
- DIFF/INCR backup has failed, FULL may be enforced for some specific task types
- This behavior depends on the task type. A failed DIFF/INCR may enforce FULL for IBM Domino (LotusNotes), IMAP, Exchange, GroupWise, KOPANO, Citrix XEN, and Hyper-V, while a failed INCR (transaction log backup) may enforce DIFF (cumulative) for MS SQL Server and SAP_ASE.
- A subtype of a VMware backup task was changed after a FULL backup
- For example, if a VMware FULL backup used the Backup as Image option (enabled by default, subtype "_IMG"), it performed a FULL raw backup of VMDK without CBT. If the subtype was changed manually, this may break the backup chain. Therefore SEP sesam enforces a full backup to create a new initial full backup saveset that is unchanged and can act as the base for subsequent differential and incremental backups. For details on the available backup options, check VMware backup.
- Exchange Server validating backup integrity by using CHKSGFILES
- CHKSGFILES is used to verify the data before performing a backup to ensure that the databases to be backed up are not corrupted. If sm_reformat_lis fails for DIFF/INCR (*.lis files were not retrieved/read), the next run of a backup job is FULL.
- A backup task is renamed or copied
- SEP sesam must be able to recognize the renamed or copied backup task as a new backup task and not as a continuation of the previously renamed task. Therefore such a newly created task is automatically executed as FULL.