Data center failures can cause data loss and service downtime. AnalyticDB for MySQL Enterprise Edition and Data Lakehouse Edition clusters support multi-zone deployment, which replicates data synchronously across two zones. When the primary zone fails, the cluster automatically fails over to the secondary zone with a recovery point objective (RPO) of 0 and a recovery time objective (RTO) of 30 minutes or less — no manual intervention required.
How it works
Multi-zone deployment is built on regional Enterprise SSDs (ESSDs), which provide native cross-zone storage redundancy.
All access nodes, compute nodes, and storage nodes run in the primary zone. Writes are committed synchronously to both zones, so the secondary zone always holds a complete, up-to-date copy of your data.
The following diagram shows the cross-zone disaster recovery architecture:
When the primary zone fails, the entire service switches over to the secondary zone automatically.
The following diagram shows the failover sequence:
Supported editions and regions
Only Enterprise Edition and Data Lakehouse Edition clusters support multi-zone deployment.
Multi-zone deployment is available in the following regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Shenzhen), China (Hong Kong), and Singapore.
Synchronous writes to two zones increase average write latency by typically within 5% compared to a single-zone cluster, based on internal testing.
Failover impact
During a primary/secondary failover, data read/write operations may fail, or read/write timeout errors may occur.
The failover completes within 30 minutes (RTO ≤ 30 minutes).
Next steps
Pricing for Enterprise Edition and Basic Edition: review multi-zone cluster pricing.
Create a cluster: create a multi-zone cluster from scratch.
Change the deployment mode: switch an existing cluster to multi-zone deployment.