Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: 现在业务的查询量要增加很多，数据量不变，如果扩容的话 应该扩容kv 还是 tidb节点?
[TiDB Usage Environment] Production Environment
[TiDB Version] v4.0.6
Environment Description:
Currently, TiDB/PD are on the same servers (3 servers, each with 256GB memory, 48 CPUs), TiKV (3 servers, each with 256GB memory, 48 CPUs, 1.5TB*3, with 3 instances, each using 1.5TB)
Currently, TiDB CPU usage is about 3%, memory usage is 5%; TiKV CPU usage is 2%, memory usage is 59%
Experts, I would like to consult: now the query volume of the business is going to increase significantly, but the data volume remains the same. If we need to scale out, should we scale out the KV or TiDB nodes?
Based on your current description, the overall resource utilization of TiDB is actually quite low. It is necessary to evaluate the specific types of SQL queries (simple point queries, small range queries, or multi-table join queries) to determine whether to scale TiDB or TiKV. TiDB is generally used to handle high concurrency, such as in a B2C business where the number of users increases from 100,000 to 10 million. TiKV, on the other hand, is used to expand data volume and throughput. Additionally, in scenarios with many complex queries, it can also increase the computational power of multiple nodes, improving the performance of aggregate and multi-table join queries.
TiDB nodes mainly address computing issues and are stateless.
TiKV nodes primarily handle storage, disk I/O, and distributed computing issues.
To determine whether to scale out either or both, you need to first assess the specific situation of business growth before proceeding with the expansion.
Speculation is useless; you need to conduct stress tests and simulate real business scenarios to see where the bottlenecks are when the load is high.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.
|
username: xiaohetao
|
username: 
|
username: 
|
username: 
|
username: