Implementation of Multi-Planet Robot Nodes

[linqingping]

How to implement network controller deployment composed of multiple planetary nodes

[zikuan-wang]

same issue

[ifyz]

same issue

I’ve been thinking about it, and if a single ztncui is managing multiple ZeroTier clients as planets, it’s still fundamentally centralized. If the server hosting the ztncui controller goes down, it would still disrupt normal usage.

A better solution might be to deploy multiple ztncui instances using the same network ID, with the controllers keeping their data in sync. But that approach might be tricky to implement.

[ifyz]

Here is a possible solution, though it has not been tested:

1. Deploy ztncui and the ZeroTier server in Docker containers on multiple servers, ensuring that all ztncui instances share the same network ID.

2. Centralize all ZeroTier servers to create a planet configuration.

3. Replace the planet file on each ZeroTier server with the configuration created in step 2.

4. When adding any network configuration, it must be seted on all ztncui panels
   (or you can write a script to manage and synchronize all ztncui configs).