Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: Tikv tidb_replica_read配置疑问
[Test Environment for TiDB]
[TiDB Version] V6.5.0
[Encountered Issue: Problem Description and Impact] Follower Read | PingCAP Documentation Center In the official documentation regarding the tidb_replica_read configuration closest-replicas, how does TiDB identify the concept of the same availability zone? Is it through the label tags of TiKV?
The Follower Read feature refers to using the follower replicas of a Region to handle data read tasks under the premise of strong consistency reads, thereby improving the throughput of the TiDB cluster and reducing the load on the leader. Follower Read includes a series of load balancing mechanisms that offload the TiKV read load from the leader replica of a Region to the follower replicas. TiKV’s Follower Read can ensure data read consistency and provide users with strongly consistent data read capabilities.
Isn’t this explained already? The capability of replicas is just like this. TiKV’s labels help you control the distribution of replicas.
For example: I need replicas to be distributed on different hosts, so even if one host fails, it won’t affect the other two hosts. Because there are replicas, you just need to add new nodes, and the replicas will be automatically scheduled… keeping the cluster healthy.
The replication capability is okay. My question is, when I configure tidb_replica_read=tidb_replica_read, how does TiDB prioritize selecting the same availability zone?
Is there a specific naming convention for the keys of labels?
Regarding the content of labels, you can refer to this:
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.
|
username: 林夕一指
|
username: 
|
username: 
|
username: 
|
username: