Issues with Using TiUniManager

Translated
1

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

Original topic: TiUniManager使用问题

| username: ddhe9527
  1. After importing the existing cluster, the maintenance status has always been “taking over,” and the restart and stop buttons are grayed out and cannot be operated. Is this status abnormal?
    Dingtalk_20220704095701
    Dingtalk_202207040957011380×422 39.1 KB
  2. After importing the existing cluster, the status of each host in resource management is abnormal. Where can I see the details of this abnormality? The numerator part of the CPU and memory is 0, such as 0/4 and 0/8. What do these represent?
    Dingtalk_20220704095646
    Dingtalk_202207040956461380×616 82.7 KB
2
| username: ddhe9527 | Original post link

Bumping this up.

3
| username: Min_Chen | Original post link

Hello, could you check if the network is smooth?

4
| username: ddhe9527 | Original post link

TiUniManager is deployed on my TiUP control machine. I have always used this control machine to manage the cluster, and there are no network issues.

5
| username: haiboumich | Original post link
  1. Does importing an existing cluster mean taking over the cluster? Could you please describe the specific steps in detail, such as on which page to import the original cluster hosts (it would be best if you could provide the imported xlsx file with sensitive parts hidden)?
  2. You can view the error information in the workflow task management. The display of CPU and memory as 0/4 and 0/8 indicates that there is no usage, but this is an abnormal state, so the data is not quite right.
6
| username: ddhe9527 | Original post link

I am taking over an existing test cluster. The specific operation is to click “Take Over Cluster” in the upper right corner of the cluster page, then upload an Excel file. The SSH user is the tidb user previously used by TiUP. Additionally, I would like to ask, if there is an issue with the takeover and I want to retry, for example, if I suspect that I entered the SSH user password incorrectly and want to troubleshoot, how should I proceed?

7
| username: TiDBer_wTKU9jv6 | Original post link
  1. What version of tiup, the latest version of tiup, in the component purpose column in Excel, Schedule is a required item.
  2. There is error information for each step in the “Workflow Task Management”.
  3. If there is an issue with the takeover, delete it directly. If it cannot be deleted, you need to redeploy tiunimanager.
  4. The takeover process requires some time, and it is normal for the button to be greyed out.
  5. “The numerator part of CPU and memory are both 0, for example, 0/4, 0/8, what does it mean?” It should be the resources allocated during the creation after rebuilding the cluster through tiunimanager, and the estimated resources for the takeover are not displayed.
8
| username: ddhe9527 | Original post link

There is a purpose column in Excel. If there is a problem with the takeover, you can only redeploy TiUniManager. This design is a bit anti-human. :joy:

9
| username: TiDBer_wTKU9jv6 | Original post link

I think the main reason is that the default value of the tidb_distsql_scan_concurrency parameter is 15. When the concurrency is high, the CPU usage will increase significantly. You can try to reduce the value of this parameter to see if it helps.

10
| username: ddhe9527 | Original post link

Aren’t Compute/Storage/Schedule corresponding to tidb/tikv/pd instances respectively? You only fill in schedule for the node where pd is deployed, isn’t that right?

11
| username: TiDBer_wTKU9jv6 | Original post link

There is a bug. Just include Storage, Schedule, and Compute, and you can choose the specifics when creating it.

12
| username: ddhe9527 | Original post link

I’ll give it a try.

13
| username: 小王同学Plus | Original post link

Is it working normally now?

14
| username: ddhe9527 | Original post link

Re-deployed TiUniManager, during the takeover process, you can see specific error information in the “Workflow Task Management”. It is temporarily normal, other functions are still being tested.

15
| username: system | Original post link

This topic was automatically closed 1 minute after the last reply. No new replies are allowed.