Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: tiup cluster display tidb-test 不能在其它机器上执行
[TiDB Usage Environment] Development
[TiDB Version]
V7.5
[Encountered Problem: Phenomenon and Impact]
The installation of tiup was executed on machine A, and the command to initialize the cluster was also executed on this machine.
When executing the command tiup cluster display tidb-test
on this machine, the entire cluster and all components on all machines are shown as up.
However, I want to use the tiup cluster display tidb-test
command on other machines, but it reports an error.
tiup is checking updates for component cluster ...
Starting component `cluster`: /root/.tiup/components/cluster/v1.14.0/tiup-cluster display tidb-test
Error: Cluster tidb-test not found
Verbose debug logs have been written to /root/.tiup/logs/tiup-cluster-debug-2023-12-13-22-43-24.log.
Other nodes also have tiup installed, all following the online installation documentation provided on the official website.
Is this result by design? The command cannot be executed on other nodes.
Yes, the tiup cluster display
command can only be run on the machine where you deployed the cluster, because the cluster’s metadata is stored in TiUP. You can copy this TiUP to other machines and use it, but it is not recommended for a single cluster to be managed by two TiUP instances simultaneously, as the metadata will not automatically synchronize.
Just deploying tiup is not enough, try referring to this:
Use tiup cluster list
to check what clusters are installed locally on this machine. If the other machine only has tiup installed, it definitely won’t display information about other clusters.
This can only be used normally on the control machine. If you want to use it on other machines, you need to install tiup first and then package the cluster metadata over.
Then make the other machines control nodes as well, but the permissions will be quite extensive, and those machines will also be able to execute the destroy operation.
Just check the PD Dashboard, it’s a graphical interface.
What kind of requirement involves operating tiup on multiple nodes? It’s quite rare and can easily mess up the cluster.
Copy the .tiup folder from other nodes and overwrite it.
Set up another control machine, but usually, it’s only done when the original control machine has issues. Your scenario is a bit unusual; can you explain what your scenario is? 
It should not be possible. Generally, tiup and cluster tools are installed on a central control machine. Although other machines also deploy cluster nodes, they may not have the corresponding installation and deployment metadata, so they cannot execute the corresponding commands.
I have the same question. If the TiUP node goes down, how can we perform maintenance work on the cluster?
Regularly back up the .tiup files and configuration files. In case of a control machine failure, deploy tiup on another server and redeploy to restore the use of tiup components.
The image you provided is not accessible. Please provide the text you need translated.
There are so many things I want to say, but I held them back… Goodbye.
It can only be done on the central control machine.
Still don’t know what the answer is…