SAP HANA HSR provides High Availability and Disaster recovery
Fundamentals of HSR:
SAP HSR has two fundamental concepts
1. HSR Replication Mode: Synchronous, Synchronous to memory, Asynchronous
2. HSR Operation Modes: Delta shipping, Log replay or log replay with read access
Replication Mode:
The HSR replication mode determines the behavior of secondary system in acquiring the data from the primary system
1. Synchronous (SYNC)
-
The primary system waits for an acknowledgment from the secondary system before committing transactions.
Ensures no data loss.
-
Best for high-availability setups within low-latency environments.
-
Requires a reliable, high-speed network.
2. Synchronous in-memory (SYNCMEM)
-
Ensures zero data loss but only guarantees that data is in the secondary system’s memory (not persisted to disk).
-
Faster than SYNC since it doesn’t wait for disk writes on the secondary system.
-
Suitable for systems with strict uptime requirements.
3. Asynchronous (ASYNC)
-
The primary system does not wait for the secondary system to acknowledge the transaction.
-
Some data loss is possible in case of failure before replication completes.
-
Best for disaster recovery setups over long distances.
4. Synchronous Full Sync (FULLSYNC) [Less Common]
-
Similar to SYNC but ensures the secondary system has fully committed the transaction before acknowledging.
-
Used in rare cases where absolute consistency is critical.
Choosing the Right Mode
-
High availability within a data center? → SYNC or SYNCMEM
-
Disaster recovery across regions? → ASYNC
-
Extreme data consistency required? → FULLSYNC
Operation Mode:
HSR Operation mode determines how the data is sent and processed by the secondary system
• SAP HANA studio
• `Hdbnsutil` (a command-line tool)
Stop the secondary system
Note: If you want to configure Active-Active Read, then you need to select the operation mode as Logreplay_readaccess.
Once the secondary system restarts automatically, the replication will begin to start.
You can monitor the replication from Hana Studio-->Landscape-->System Replication.
Step 3 : Takeover
- You can trigger the takeover in the event of disaster or testing.
- Register the former primary system as a new secondary system.
Once the former primary comes back online, you need to register it as the secondary to the former secondary.
5. Check the details you have provided in the below screen and click on Finish to complete the task.
No comments:
Post a Comment