TiDB Cloud Deployment

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

Original topic: TiDB云端部署

| username: HACK

If TiDB is deployed on the enterprise’s internal cloud, should the server be a virtual machine or a physical server? Are there any specific practical experiences with this?

【TiDB Usage Environment】Production Environment / Testing / PoC
【TiDB Version】
【Reproduction Path】What operations were performed that led to the issue
【Encountered Issues: Issue Phenomenon and Impact】
【Resource Configuration】
【Attachments: Screenshots/Logs/Monitoring】

| username: TiDBer_jYQINSnf | Original post link

If the DB is deployed on the internal cloud of the enterprise, should the server be a virtual machine or a physical server? Are there any specific practical experiences?

K8s manages physical servers + local disks.
It’s unclear what vendor your internal cloud is from, and it also involves K8S, network integration, local disk allocation plugins, etc. Not sure to what extent your cloud has been developed.

| username: tidb菜鸟一只 | Original post link

For enterprise cloud using TiDB, it generally involves k8s + cloud disks. It would definitely be better if multiple physical servers could be deployed.

| username: HACK | Original post link

How do you manage TiDB with k8s? Do you have to use k8s?

| username: ohammer | Original post link

You can check this out, there are specialized tools. TiDB on Kubernetes 用户文档 | PingCAP 文档中心

| username: tidb菜鸟一只 | Original post link

For enterprise cloud, if you can get multiple high-performance physical servers, you can definitely deploy TiDB, and the performance will be better than deploying on k8s. It is also simpler and easier to maintain. However, you need to pay attention to the network conditions between the physical servers. The key point is that if it is enterprise cloud, they generally won’t provide you with multiple physical servers because the cost is much higher. Usually, it is a set of k8s plus cloud disks, which is relatively low cost.

| username: TiDBer_jYQINSnf | Original post link

This is deployed on k8s. There is no performance loss for physical machines managed by k8s, but you need to consider network connectivity, disk allocation, and also understand k8s. It’s not something that can be done in a day or two. Even if you can deploy it step by step, it will be difficult to use if you are not familiar with k8s operations later on. If you understand k8s, it will be much simpler.

| username: system | Original post link

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