The primary/standby cluster architecture plays an important role in ensuring high availability (HA) of OceanBase Database, and provides more flexible disaster recovery and HA solutions.
This topic describes the benefits of this architecture.
Cluster independence and fault isolation
Compared with the architecture of a single cluster with multiple replicas, the primary and standby clusters are physically independent of each other. A failure in one cluster does not affect other clusters.
Flexible HA
In the case of planned or unexpected downtime of the primary cluster, a standby cluster can become a new primary cluster through switchover or failover. Dual-IDC deployment and disaster recovery across two regions are supported.
Flexible data protection
The maximum protection, maximum performance, and maximum availability modes are provided to help users achieve a balance between data availability and performance.
Automatic data synchronization
The primary cluster proactively sends REDO logs to the standby clusters to ensure consistency. If the network connection between the primary cluster and a standby cluster is interrupted, the logs on the standby cluster are out-of-sync with the primary cluster. After the network connection resumes, the standby clusters proactively synchronize REDO logs from the primary cluster without manual intervention.
Physical data consistency
The data and schemas are physically consistent between the primary and standby clusters. The primary and standby clusters perform checksum verification to ensure data consistency.
Real-time data replay
The standby clusters can replay REDO logs in real time and accept external read requests. Business that does not have high requirements for real-time performance, such as report generation and statistics collection, can be placed in the standby clusters to reduce the load of the primary cluster and efficiently use CPU and I/O resources.
Heterogeneous deployment and separate O&M
Different resource specifications, including the CPU resources, memory resources, and replica quota, can be configured for the primary and standby clusters. In addition, the primary and standby clusters have separate O&M processes.