How to easily back up the SAP HANA database?

In 2010, SAP released a product whose full name is SAP High Performance Analytic Application, which is known as SAP HANA. After years of development, SAP HANA has evolved from an "application" to a high-performance application platform with built-in computing, database, analysis and forecasting engines. Now more and more modules are gradually moving closer to or integrated with this platform.

In order to ensure its own best performance, SAP HANA stores data in memory, and at the same time, uses a persistent storage system for fault recovery. When the database is in normal operation, data and undo logs are automatically persisted to the hard disk during the savepoint process, and data changes are recorded in the redo log. Although savepoints and log write operations can prevent unexpected factors such as power outages and hardware failures from affecting the database, they are powerless when the persistent storage device fails or is misoperated. In order to prevent data loss caused by the above situations, the backup of the SAP HANA database is absolutely necessary.

1. SAP HANA backup features:
SAP HANA backup can be divided into log backup and data backup. The specific features include:

1) The backup process of the two parts is independent of each other. Before SAP HANA performs the first data backup, the log backup will not be performed;
2) Backup and recovery are applied to the entire database, and a certain object in the database cannot be backed up and restored;
3) SAP HANA backup requires at least BACKUP ADMIN and CATALOG READ permissions;

2. Technical implementation of Alibaba Cloud Hybrid Cloud Backup (HBR)
Alibaba Cloud Hybrid Backup (HBR) launched the SAP HANA BACKINT integration solution a year ago, becoming the first cloud vendor in Asia to pass the SAP HANA BACKINT certification. In addition, HBR can also provide Alibaba Cloud ECS, NAS, OSS and self-built Files in the computer room, VMware, SQL Server, etc. provide safe and efficient backup and disaster recovery protection.

The traditional backup solution is shown in the figure below. It needs to rely on tools such as SAP HANA Studio or cockpit to dump the data to be backed up to the local disk or NAS, and then upload it to OSS. It can be seen that the configuration in the backup solution is very cumbersome. It also needs to be matched with various scripts for life cycle management.

2.jpg

In order to simplify this process, HBR integrates the SAP HANA BACKINT interface to perform three-party integration with SAP HANA. When SAP HANA starts the backup process, it will start the plug-in deployed by HBR to back up the data to the HBR backup database. HBR will participate in all scripts required The steps are fully automated management.
In addition, the HBR HANA plug-in follows the SAP HANA Backint interface specification, and the SAP HANA database stores streaming data in the HBR HANA plug-in, and the restore is initiated by SAP HANA, and then the HBR plug-in plays back the data stream during backup.

3.jpg

Through the above two technical means, the backup efficiency of SAP HANA is greatly improved. At the same time, in order to achieve faster data recovery, SAP HANA usually adopts the method of full backup + log backup, which leads to how to reduce the storage of full backup Space occupation. In order to solve this problem, HBR adopts a variable-length deduplication algorithm after the streaming protocol. The variable-length deduplication algorithm can better limit the impact of data changes to one slice or between two adjacent slices. Such an algorithm achieves a deduplication ratio of 30:1, which can effectively reduce the storage capacity of SAP HANA full backups and effectively reduce storage costs for users.

3. Three methods of HBR backup SAP HANA
There are currently three deployment methods for SAP HANA:
1. SAP HANA single node deployment

4.jpg

-The HBR console sends control commands directly to the HBR agent (hybridbackup) on the SAP HANA node
-Hybridbackup generates corresponding SQL statements according to the control commands and sends them to the SAP HANA database.
-SAP HANA database starts the backup task after receiving the corresponding SQL statement and calls BACKINT to read the backup data.
-BACKINT redirects the data stream to IDS for deduplication and compression processing, and finally saves it in the HBR backup library.

2. SAP HANA dual-machine HA deployment - divided into primary node and standby node, if the primary node fails, the primary node will be switched to the standby node

5.jpg

-HBR needs to deploy HBR Agent on two SAP HANA nodes, but only one of the data plane programs (BACKINT and IDS) working at the same time is consistent with the SAP HANA master node.
- Control commands The HBR console randomly selects an online hybrid backup to send control commands.
-hybridbackup generates SQL commands to send the commands to the high-availability virtual IP address, and the high-availability virtual IP address will route the request to the current active node for execution

3. SAP HANA Scale-out deployment - usually HBR backup SAP HANA scale-out for database deployment with a large amount of data

6.jpg

-SAP HANA Scale-out deployment adopts the Master-worker structure, there is a master node, and all backup and recovery commands need to be sent to the master node for execution
-SAP HANA workers also generally distribute some data, so when the backup is executed, the corresponding master and worker will generate data streams.
-The control command is still sent by the HBR console to arbitrarily select the online hybrid backup node.

4. Advantages of HBR backup SAP HANA
Compared with backing up the SAP HANA database locally, the advantages of backing up the SAP HANA database on Alibaba Cloud include:
1. High reliability: the backup end uses multiple copies for storage, thereby achieving 99.999999999% data reliability;
2. More secure: Multi-user resource isolation mechanism, support remote disaster recovery mechanism, provide multiple authentication and authorization mechanisms, white list, anti-leeching, master and sub-account functions, etc.;
3. More economical: the backup library adopts deduplication storage, supports various payment forms such as pay-as-you-go, resource packs, etc., and reduces the overall cost of ownership;
4. Easier: The operation is simple and easy to use, and the periodic backup of the SAP HANA database and related configurations of log backup can be easily realized through the HBR console interface.