Performing Database Recovery

Hardware failures or user errors are the main reasons that might require a database restore, and are explained in the following list:

(1) Data volume is unusable

If the data area is unusable, and all the data changes after the last complete data backup are still available in the log backups and log area, you can recover the data from committed transactions that was in-memory at the time of failure.

No committed data is lost.

For recovery, the data backups, the log backups, and the log area are used. When you restore the data backup successfully, the log entries from the log backups and the log area are replayed automatically.

You can also recover the database using an older data backup and log backups. All relevant log backups made after the data backup are needed for the recovery.

For more information, see SAP Note 1705945: Determining the files needed for a recovery.

(2) Log volume is unusable

If the log area is unusable, you can only replay the log backups. Therefore, any changes that are made after the most recent log backup are lost. In addition, all the transactions that were open during the log backup are rolled back.

You can still recover the database to a point in time within the existing log backups.

For recovery, the data backups and the log backups are used. When the data backup is successfully restored, the log entries from the log backups are automatically replayed. To prevent the recovery of entries from the unusable log area, in the Recovery Wizard, specify the Initialize log area option.

(3) Logical Error: Point in Time recovery

To reset the database to a particular point in time, you need a data backup from before the point in time to recover to, the subsequent log backups, and the log area.

All changes made after the recovery time are lost. If you perform this recovery, consider recovering the database to a different system.

When a database has crashed due to hardware errors of user error, the database administrator need to restore the database to a consistent state.

The following recovery types are available:

(A) Recover the database to its most recent state

This option recovers the database to as close as possible to the current time.

This recovery option uses the following data:

• The most recent data backup

• Log backups made since the most recent data backup

• Log area

(B) Recover the database to a specific point in time

This recovery option uses the following data:

• The last data backup available before the specified point in time

• Log backups made since the data backup to be used

• Log area

(C) Recover the database to a specific data backup

This recovery option uses the following data:

The specified data backup

• SAP HANA database must be shut down

• User must have the system privilege BACKUP ADMIN

• The <sid>adm operating system user is required

• At least one data backup is required

• Before a recovery can start, all data and log backups must be either accessible in the file system or available through the third-party backup tool

• The number and type of services in source and target systems must be identical

• To recover customer-specific configuration settings, configure the customer-specific settings before starting the recovery

A recovery typically consists of the following phases:

1. Data recovery - complete data backup + delta backups
2. Log replay - log backups + log entries from the log area
3. Restart

To perform an SAP HANA database recovery, the following requirements must be met:

• The SAP HANA database software must be installed, so that an initial database exists. In a recovery situation, you can use the SAP HANA cockpit to restore customer-specific changes to this initial database.

  Note that if you want to restore customer-specific configuration settings, you can do this either before you restore the database and the log backups, or at the end of the recovery.

• Ensure that the target system and the source system have identical configurations. The number and types of services (for example, index server) on each host must be identical for both system landscapes.

• At the beginning of a recovery, all the data and log backups to be used must be either accessible in the file system or available through the third-party backup tool.

• At least one data backup must be available before the recovery is started.

• To restore the database to a particular point in time, you need a data backup and all the log backups up to the point in time for recovery (including the log backups made after the desired point in time of the recovery).

• The SAP HANA database software is installed on the target system.

• To perform a recovery, an SAP HANA database needs to be shut down.

• The following must be available:
  • At least one full backup (complete data backup or data snapshot)
  • If required, delta backups created since the full backup to be used

  • If required, log backups created since the full backup to be used

    (Covering changes not already contained in the delta backups)

  • If required, the log area

• The recovery of a tenant database is always initiated from the system database. If tenant databases need to be recovered, they are recovered individually, and not all together in one single operation.

• The system database only needs to be recovered if it is corrupted. If only a tenant database is corrupted, the system database does not need to be recovered.

• If the system database is shut down for recovery, all its tenant databases are automatically shut down as well.

• When an SAP HANA multitenant database container is recovered, the services needed are generated automatically in the tenant databases.

• To recover a complete SAP HANA system, the system database needs to be recovered first, and then all the tenant databases are recovered individually.

• Tenant databases are recovered individually, and not all together in one single operation.

• Recovery of tenant databases can only be initiated from the system database.

• You need the logon credentials of the system database user.

• The system database and other tenant databases are not affected and remain online.

• The whole system is shut down, including all tenant databases.

• You need the logon credentials of the <sid>adm user.

• After the system database is recovered and restarted, you can restart the tenant databases.

• The content of the tenant databases is not affected by the system database recovery.

The options to recover a database using SAP HANA cockpit are as follows:

• To its most recent consistent state

• To a specific point in time