Concepts Related to HA and Disaster Recovery

• Availability

Availability is used to measure the running continuity of the system. It is expressed in percentage and is inversely proportional to the downtime. For example, if the availability of a system is 99.9%, its annual downtime must be less than 0.1% or 9 hours.

• Downtime

Downtime is the result of service interruption, which may be caused by system upgrade or accident faults. A fault may be a device, software, or network fault, or a major disaster, such as a fire. A power failure or construction accident may cause the failure of the entire data center.

• HA

HA is ensured by a set of business technologies, engineering practices, and design principles. System continuity can be achieved by eliminating a single point of failure. HA minimizes the service loss by restoring the system immediately after it is interrupted.

• DR

Disaster recovery is performed after a data center is interrupted or onsite faults are being rectified. Data backups prepared in event of disasters are more complex and expensive.

• Data synchronization mode
  The SAP HANA System Replication supports synchronization (full synchronization, memory synchronization, and synchronization) and asynchronization modes. In the HA automatic switchover scenario, the synchronization mode is recommended to ensure that no data is lost during the data switchover (RPO = 0).

  • Full synchronization: After the secondary system receives the synchronization data and stores it in disks, the secondary system sends confirmation information to the primary system and then the primary system submits logs of this operation. If the secondary system fails to receive the synchronization data due to a fault, the primary system waits until the secondary system recovers.
  • Synchronization: After the secondary system receives the synchronization data and stores it in disks, the secondary system sends confirmation information to the primary system and then the primary system submits logs of this operation. If the secondary system fails to receive synchronization data due to a fault and the primary system fails to synchronize data to the secondary system, the primary system continues its service.
  • Memory synchronization: After the secondary system receives the logs (memory) and sends a confirmation message to the primary system, the primary system submits logs of this operation. If the secondary system fails to receive synchronization data due to a fault and the primary system fails to synchronize data to the secondary system, the primary system continues its service.
  • Asynchronization: The primary system sends the synchronization data to the secondary system and submits logs without waiting for the response from the secondary system. The fault of the secondary system does not affect the primary system running services.
• Data preload mode
  • The preload option is ON.

    Data is stored in the memory.

    The switchover is fast (RTO).

  • The preload option is OFF.

    The memory usage is low and the memory is used by other systems, such as non-production systems.

    The switchover duration is long.

    Before the switchover, activate the preload mode.