Is it possible to improve performance by deploying two TiKV nodes on a single disk?

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

Original topic: 一块磁盘上部署两个tikv 节点有可能提高性能吗?

| username: 数据库菜鸡

Will deploying multiple TiKV nodes on a single NVMe disk improve overall performance if there are currently only three machines, each with one NVMe disk?

| username: 像风一样的男子 | Original post link

Creating multiple TiKV instances on a single SSD will definitely result in poorer performance compared to having one SSD per TiKV instance. It is recommended to have one NVMe per TiKV node. You can insert multiple SSDs into a single physical server, create multiple RAID groups, and create one TiKV instance per RAID group.

| username: h5n1 | Original post link

The more a resource is used, the more contention there will be, and performance will decline.

| username: tidb菜鸟一只 | Original post link

How about deploying a TiKV node on a single disk?

| username: 啦啦啦啦啦 | Original post link

If there are two disks, you can consider a hybrid deployment.

| username: songxuecheng | Original post link

Feasible when the IO requirements are not high.

| username: zhanggame1 | Original post link

I feel like it will reduce performance. Is there any reason to believe it will improve performance?

| username: cy6301567 | Original post link

IO competition will lead to performance degradation, so try to deploy KV nodes and PD separately.

| username: system | Original post link

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.