Restore and recover SAP HANA Scale-Out instances

This page includes information on recovering SAP HANA Scale-Out instances from Backup and DR Service backups.

License requirements and impact on restore

The license key for an SAP HANA database is based on the system ID and the hardware ID. After a recovery, an SAP HANA license key becomes invalid if the SID or hardware ID has changed.
During recovery, a temporary license key is installed automatically if the backup used for recovery has a permanent license, which is still valid. You can work with the automatically installed temporary license for up to 90 days. During this time, you need to apply to SAP to have the license from the source database transferred to a new license key. You then need to install the new license key in the recovered SAP HANA database.
If the backup that was used for recovery only had a temporary license, the database is in lockdown mode immediately after recovery.

Source database with temporary license These are backups taken with temporary licenses.

Restore back to source.: It is 90 days from the time of database creation and the database is in lockdown mode.
Restore to the new target. It fails as SAP temp license does not allow the restore to new target.

Source database with permanent license These are backups taken with permanent licenses.

Restore back to source. No issue.
Restore to the new target It has 90 days trial license. Backups succeed but they are not be able to use the backup to restore.

SAP references

Preflight check

During the restore procedure, a preflight checks validate the recovery. These required prerequisites are checked for a successful database restore:

HANA SID: HANA is configured on the target node with the same HANA SID name.
Config file: Config file global.ini is configured correctly
Log backup path: Log backup path is set under global.ini
BACKINT CONFIG:
- From volume-level image: Backint is not configured for this database.
- From full+incremental image: If the target server is not already configured with backint, then backint is configured during restore.
HANA VERSION: Target HANA version is the same as the source HANA version.
USERSTOREKEY: Provide userstore key exists on target or a valid privileged username and password is specified which exists at the time of backup.
Logical volume
- The logical volume name and volume group name should be the same on source and target node.
- The logical volume size on target should be the same or greater than source logical volume.
Node Status
- Scale-up and standalone config: Target node is up and available
- HANA HA (1+n) config: Replication needs to disabled before restore can be initiated. Post restore replication needs to be reconfigured. Restore to any node of a HANA HA (1+n) cluster results in creating a standalone application on the selected target host. User has to explicitly configure the cluster back as needed and discover the application appropriately

Automated recovery of an SAP HANA database

Before you begin

Be sure to review License requirements and impact on restore
Be sure to review Prerequisites required for the successful database restore

To new from volume-level backup image

Automated recovery of a HANA database to a new target from a volume-level backup image

From the App Manager Applications list, right-click the database and select Access.
Select the latest snapshot to recover and choose Restore.
On the Restore page, choose Restore Back to New Target.
- Target. For all configurations, eligible HANA nodes will be available to choose under the drop down. Select the node for restore from the drop down.
- Replace Original Application identity. This option is only available when restore is performed to a new host on the same appliance where the backup was originally generated.
  - Yes. This will replace the original application and will carry the same application ID, jobhistory, backup images and backup plan as the original application.
  - No. This won't replace the original application. It will be discovered as a new application as part of restore job.
Note: After a HANA HA restore, the node will become standalone and retain the same appid of the cluster. If you enable replication, the next discovery on the restore host will find the cluster and continue to use the same appid as cluster host.
- Rollforward time. Choose a date and time for a database protected with logs to recover to.
- TARGET DATABASE SID. This will be pre-populated with the protected database sid name and is immutable.
- SAP DB USER STORE-KEY. This will be pre-populated with the user store key during the backup. A new userstore key can be specified with a privileged username and password that was available during backup. This new userstore key will be created and will be used for recovery.
- If the username, password is provided with the existing userstore key, the userstore key will be recreated with this username and password. The validation will only be done after the systemdb recovers. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
- If a new userstore key with username and password is specified, then the userstore key will be created with the specified userstore key name and username and password. The validation will only be done after the systemdb recovery. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege or is not available as part of the backup image.
- If no userstorekey, username and password are passed, then during the precheck, validation will occur to check if the userstorekey used during backup exists on the target server. The pre check will fail if the userstorekey used during backup is not found. Tenant recovery may fail if the username or password is not valid or or does not contain the right privilege or is not available as part of the backup image.
- If no userstorekey is passed, but username and password are provided, then the userstorekey used during backup will recreated with the credentials passed on the target server. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
Click the pre-flight check.

If the pre-flight check fails, fix the issue and resubmit the Pre-Flight Check.
If the pre-flight check is successful, click Submit to submit the restore job.

To source from volume-level backup image

Automated recovery of a HANA database to the source from a volume-level backup image

From the App Manager, Applications list, right-click the database and select Access. From the latest snapshot to recover, choose Restore.
On the Restore page, choose Restore Back to Source.
- Target. complete the following:
  - For standalone SAP HANA configuration, Target is pre-populated.
  - For HANA HA (1+n) configuration, select the HANA HA node to restore to from the drop-down list.
- Rollforward time. Choose a date and time for a database protected with logs to recover to.
- TARGET DATABASE SID. This will be pre-populated with the protected database sid name and is immutable.
- SAP DB USER STORE-KEY. This will be pre-populated with the user store key during the backup. A new userstore key can be specified with a privileged username and password that was available during backup. This new userstore key will be created and will be used for recovery.
  - If the username and password are provided with the existing userstore key, the userstore key will be recreated with this username and password. The validation will only be done after the systemdb recovers. The tenant recovery may fail if the username or password is not valid or does not contain the right privilege and or not available as part of the backup image.
  - If a new userstore key with username and password is specified, the userstore key will be created with the specified userstore key name and username and password. The validation will only be done after the systemdb recovery. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If no userstorekey, username and password are passed, then during the pre-check, validation will occur to check if the userstorekey used during backup exists on the target server. The pre-check will fail if the userstorekey used during backup is not found. The tenant recovery may fail if the credentials are not valid or or don't contain the right privilege and or are not available as part of the backup image.
  - If no userstorekey passed, but username and password was provided, then the userstorekey used during the backup will recreated with the credentials passed on the target server. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
1. Click Pre-Flight Check.
- If pre-flight check fails, fix the issue and resubmit the Pre-Flight Check.
- If the pre-flight check is successful, then click Submit to submit the restore job.

To new from full+incremental backup image

Automated recovery of a HANA database to a new target from a full+incremental backup image

From the management console App Manager, Applications list, right-click the database and select Access.
Select the latest snapshot to recover, and choose Restore.
On the Restore page, choose Restore to a New Target.
- Target. For standalone HANA configuration this is pre-populated. HANA HA nodes will be available to choose under the drop down; select the node for restore from the drop down.
- Replace Original Identity. This option is only available when restore is performed to a new host on the same backup/recovery appliance where the backup was originally generated.
  - Yes. This will replace the original application and will carry the same application ID, job history, backup images, and backup plan as the original application.
  - No. This won't replace the original application. This will be discovered as a new application as part of the restore job. After a HANA HA restore, the node will become standalone and retain the same appid of the cluster. If you enable replication, the next discovery will find the cluster and continue to use the same appid as cluster host.
- INCLUDE LIST. For recovering SYSTEMDB with or without or one or more tenant databases out of n tenant databases, provide a comma separated list of databases under INCLUDE.
- For EXCLUDE LIST, for excluding SYSTEMDB or one or more tenant database during recovery out of n tenant database: provide a comma-separated list of databases under EXCLUDE. For example, putting SYSTEMDB in the exclude list excludes SYSTEMDB from recovery and recover all tenant databases that are backed up.
  
  Note: Use either INCLUDE LIST or EXCLUDE LIST for specific database recovery. For example if you have SYSTEMDB and tenant—tn1, tn2, tn3—then complete one of the following:
  - If INCLUDE LIST and EXCLUDE LIST are both empty, then SYSTEMDB and all tenants (tn1,tn2,tn3) will be recovered.
  - If you want to recover a single tenant tn1, then use the include list with tn1.
  - If you want to recover tn2 and tn3 and exclude SYSTEMDB and tn1, then use the include list with only tn2 and tn3.
  - If you want to recover all tenants (tn1,tn2,tn3) only and exclude SYSTEMDB, then either exclude SYSTEMDB or include tn1, tn2, tn3.
  Note: SYSTEMDB recovery impact: With SYSTEMDB recovery, any tenant db excluded from the recovery won't start after the restore. The excluded tenant will have to restarted manually.
- Rollforward time. Choose a date and time for a database protected with logs to recover to.
- TARGET DATABASE SID. This will be pre-populated with the protected database sid name and is immutable.
- SAP DB USERSTORE KEY. This will be pre-populated with the user store key during the backup. A new userstore key can be specified with a privileged username and password that was available during backup. This new userstore key will be created and will be used for recovery.
  
  Important: Tenant name on the target server should be same as source. Data volume locations for the tenant should be the same as source.
  - If the username, password is provided with the existing userstore key, the userstore key will be recreated with this username and password. The validation will only be done after the SYSTEMDB recovers. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If a new userstore key with username and password is specified, the userstore key will be created with the specified userstore key name and username and password. The validation will only be done after the SYSTEMDB recovery. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If no userstorekey is specified, then the username and password are passed, then during the precheck, validation will occur to check if the userstorekey used during backup exists on the target server. The precheck will fail if the userstorekey used during backup is not found. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If no userstorekey was passed, but username and password was provided, then the userstorekey used during backup will recreated with the credentials passed on the target server. The tenant recovery may fail if the username or password is not valid, if it does not contain the right privileges, or if it is not available as part of the backup image.
Click the Pre-Flight Check.
- If the pre-flight check fails, fix the issue and resubmit the Pre-Flight Check.
- If the pre-flight check is successful, then click Submit to submit the restore job.

To source from a full+incremental backup image

Automated recovery of a HANA database back to the source from a full+incremental backup image

From the management console App Manager, Applications list, right-click the database and select Access.
Select the latest snapshot to recover, and choose Restore.
On the Restore page, choose Restore Back to Source.
- For Target, complete the following:
  - For standalone HANA configuration this is pre-populated.
  - For HANA HA (1+n) configuration, HANA HA nodes will be available to choose under drop down. Select the node for restore from the drop down.
- For INCLUDE LIST, for recovering SYSTEMDB or one or more tenant database out of n tenant database, provide a comma separated list of database under INCLUDE.
- For EXCLUDE LIST, for excluding SYSTEMDB or one or more tenant database during recovery out of n tenant database: provide a comma-separated list of databases under EXCLUDE. For example, putting SYSTEMDB in the exclude list excludes SYSTEMDB from recovery and recover all tenant databases that are backed up.
  
  Note: Use either INCLUDE LIST or EXCLUDE LIST for specific database recovery. For example if you have SYSTEMDB and tenant—tn1, tn2, tn3—then complete one of the following:
  - If INCLUDE LIST and EXCLUDE LIST are both empty, then SYSTEMDB and all tenants (tn1,tn2,tn3) will be recovered.
  - If you want to recover a single tenant tn1, then use the include list with tn1.
  - If you want to recover tn2 and tn3 and exclude SYSTEMDB and tn1, then use the include list with only tn2 and tn3.
  - If you want to recover all tenants (tn1,tn2,tn3) only and exclude SYSTEMDB, then either exclude SYSTEMDB or include tn1, tn2, tn3.
  Note: SYSTEMDB recovery impact: With SYSTEMDB recovery, any tenant db excluded from the recovery won't start after the restore. The excluded tenant will have to restarted manually.
- For Rollforward time, choose a date and time for a database protected with logs to recover to.
- For TARGET DATABASE SID, this will be pre-populated with the protected database sid name and is immutable.
- For SAP DB USER STORE-KEY, this will be pre-populated with the userstore key during the backup. A new userstore key can be specified with a privileged username and password that was available during backup. This new userstore key will be created and will be used for recovery.
  - If the username and password are provided with the existing userstore key, the userstore key will be recreated with this username and password. The validation will only be done after the systemdb recovers. The tenant recovery may fail if the username or password is not valid or does not contain the right privilege or is not available as part of the backup image.
  - If a new userstore key with username and password is specified, the userstore key will be created with the specified userstore key name and username and password. The validation will only be done after the systemdb recovery. The tenant recovery may failif the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If no userstore key, username and password are passed, then during the precheck, validation will occur to check if the userstorekey used during backup exists on the target server. The pre check will fail if the userstorekey used during backup is not found. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - If no userstorekey is passed, but username and password are provided, then the userstorekey used during backup will recreated with the credentials passed on the target server. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
  - With SYSTEMDB recovery, the hdbuserstore key is validated at the end of SYSTEMDB recovery and before starting tenant recovery. The tenant recovery may fail if the username or password is not valid or or does not contain the right privilege and or not available as part of the backup image.
Click Pre-Flight Check.
- If pre-flight check fails then fix the issue and resubmit the Pre-Flight Check.
- If the pre-flight check is successful, then click Submit to submit the restore job.

Manual recovery of an SAP HANA tenant database

You can manually recover a a single tenant database back to the source from a volume-level backup image.

Procedure

To recover a single-tenant database, complete the following:

From the App Manager, Applications list, right-click the database and select Access.
From the runway, select the latest snapshot to recover, and then select Mount on the right.
On the Mount page disable Application Options, and under Mount Options set a mount location such as /testmnt.
Log into the server as root and change the directory to /act/custom_apps/saphana/restore: cd /act/custom_apps/saphana/restore
Execute the script for recovery:

      ./CALL_LVM_single_tenant_recover.sh DBSID TENANT SID SYSTEMDB USERSTORE KEY RECOVERY TIME

Description of arguments to the script:

    DBSID = The database SID
    TENANT SID = The name of the tenant database to be restored
    SYSTEM DB USERSTORE KEY = The System DB userstore key
    RECOVERY TIME = The required recovery time in the format YYYY-MM-DD HH24:MI:SS in UTC

For example: ./CALL_LVM_single_tenant_recover.sh lv1 lv1 ACTBACKUP '2019-09-24 20:00:00'

Once the script has completed, the Tenant DB is recovered to the point in time and is available for access.
Go to the management console and unmount and delete the backup image.

Review the status of your backups in SAP HANA Studio

You can review backup status, metadata, and backup images in HANA Studio, but you cannot access the Backup and DR Service backup images from HANA Studio. You must access backup images for recovery or other uses from the management console.

In HANA Studio, go to the Backup folder.
Go to the Backup catalog tab.

The Backup catalog tab displays the status of your backups and details such as start time, duration, size, backup type, and destination.

HANA and HANA HA 1+n restore behavior

This section includes information about HANA and HANA HA 1+n restore behaviors.

All HANA configurations

Restoring to a new target with Manage New Application option enabled:

The restored application are protected, but the Application Details & Settings section has only the default values. You must manually set the required values under Application Details & Settings after restore for the backup to succeed properly.
The restored application protection will be in a disabled state. You must enable protection from App Manager > Applications > Manage Backup plan for backups to start running.

HANA HA (1+n) configurations

For a HANA 1+1 cluster with Node A primary and Node B secondary—Node A —-> Node B—the cluster is discovered as an application under Backup and DR Service and backup runs from the Node A—primary.

Restore to a new target HANA database—standalone or cluster

If a new target is discovered as an application or protected then before restore, this application needs to be unprotected and deleted from Backup and DR. The job fails if the new target application exists.

Restore to Node A primary

Node B needs to be unregistered from the cluster before running the restore to Node A.
Node A becomes a standalone application post restore to Node A.
Node B must be registered to Node A with Node B as the secondary and a force discovery must be run on Node A to continue as a cluster backup.

Restore to Node B secondary

Node B must be unregistered or else a take-over must be run on Node B before restoring to Node B.
Shut down Node A before restoring to Node B, and add Node A as secondary to Node B post restore.
Node B is a standalone application after the restore operation. Node A must be registered to Node B with Node A as the secondary.
Run a discovery from Node B to re-discover it as a cluster application.

Note: Attaching Node B as secondary to Node A will keep the original data of A and wipes the Node B restored data. To keep Node A and Node B as protected standalone applications, run discovery on Node A and Node B after the restore without enabling the replication.
To keep Node A and Node B as standalone applications and protected, run discovery on Node A and Node B after the restore without enabling the replication.

How to disable SAP HANA system replication for restore

Log onto both systems as the operating system user—user adm.
Stop the secondary system: sapcontrol –nr -function StopSystem HDB.
On secondary system unregister the secondary system: hdbnsutil -sr_unregister --id=(secondarySiteID).
Disable system replication on the primary system: hdbnsutil –sr_disable.
Check system replication, with one of the following scripts:
- The systemReplicationStatus.py script. This script shows replication status and database information
- The hdbnsutil -sr_state script. This script shows the replication nodes role and host mapping details.

Backup and DR Service documentation for SAP HANA Scale-Out

This page is one in a series of pages specific to protecting and recovering SAP HANA Scale-Out instances with Backup and DR Service. You can find additional information in the following pages: