Dear teachers, I would like to ask how to deploy a TiDB cluster. We need a production cluster, a local cluster, and a disaster recovery cluster. The plan is to use TiCDC to synchronize data between different clusters. The idea is to synchronize the production cluster’s data to the local cluster and the disaster recovery cluster using TiCDC.
In this case, how should TiCDC be deployed? Additionally, if we deploy TiCDC in the production cluster, can the production cluster’s TiCDC create two tasks to synchronize data to both the local cluster and the disaster recovery cluster?
Why not directly use the cluster topology of the third center and use TiCDC instead?
If you use TiCDC, just configure two tasks: one source end is configured with the production cluster, the target end is configured with the same city cluster, and the other source end is configured with the production cluster, the target end is configured with the disaster recovery cluster.
It’s better to use cross-data center deployment clusters.
Why not directly use the cluster topology of the third center and use TiCDC?
If you use TiCDC, just configure two tasks: one source end with the production cluster and the target end with the same city cluster, and the other source end with the production cluster and the target end with the disaster recovery cluster.
Or use cross-data center deployment clusters.
Why not directly use the cluster topology of the third center and use TiCDC?
What does this mean? Different clusters can only use TiCDC or TiDB Binlog for different purposes, and TiDB Binlog is not recommended.
The main issue is network latency. TiCDC is not recommended to be deployed across data centers. It is suggested to deploy TiCDC downstream to reduce latency issues caused by network delays. If TiCDC is deployed downstream, it is recommended to keep the network latency between TiCDC and upstream around 50ms, and it should not exceed 200ms at most.
|
username: Raymond
|
username: 
|
username: 
|
username: 
|
username: