Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: tidb server cpu分布不均衡
How is resource scheduling handled when TiDB is computing? Currently, after submitting SQL, I found that only the CPU usage of the TiDB server node where the SQL was submitted is close to 90%, while the usage of other nodes is very low. Can anyone share their experience?
What is used for load balancing between TiDB servers? TiDB itself does not have scheduling between TiDB servers. For example, if a large SQL or multiple large SQLs are running on a particular server, you will see the situation you described.
Does it mean that when a single SQL requires a lot of resources, it is not possible to schedule resources from other nodes?
Adding load balancing on top of the TiDB server can only address the request pressure on the cluster from multiple requests, but what about when a single request has too much pressure?
No, the TiDB server is stateless, does not store data, and does not share it.
You can use several instances for OLTP applications and separate instances for OLAP applications, so they do not affect each other; and the separated instances can also be roughly balanced.
Are you referring to the TiDB server?
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.