Issues Related to High Availability Architecture in Dual Data Centers in the Same City

Translated
1

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 同城两中心高可用架构相关问题

| username: Raymond

Dear teachers, according to the TiDB official course on the two-center architecture in the same city, the region has 5 replicas. The course mentions that when the main center goes down, in synchronous mode, the disaster recovery center requires manual intervention. Could you please explain why manual intervention is needed? According to my understanding:

  1. If the main center has 3 regions and the disaster recovery center has 2 regions, when the main center goes down, 3 out of 5 regions are down, making the region unusable. This is understandable and requires manual intervention.

  2. However, if the main center has only 2 regions and the disaster recovery center has 3 regions, even if the main center goes down, the regions in the secondary center can still be used, satisfying the majority principle of Raft.

2
| username: ddhe9527 | Original post link

Uh, I suggest reading the official documentation on adaptive synchronization for two centers in the same city a couple more times to understand what synchronous mode and asynchronous mode are.

3
| username: cs58_dba | Original post link

For local disaster recovery, it’s common to use dark fiber and require synchronous replication. For remote disaster recovery, due to the long distance and reliance on the internet, only asynchronous replication can be required.

4
| username: system | Original post link

This topic was automatically closed 1 minute after the last reply. No new replies are allowed.