Skip to content

[feature] use try mode when applying config to machines #1928

@utkuozdemir

Description

@utkuozdemir

When we apply config to machines, we can use the try mode - if anything breaks, e.g., networking, the machine comes back that way.

If the machine is healthy after the try mode apply for a while (e.g., 50 seconds, less than a minute before it reverts automatically), we can do the real apply.

Note: we need to check the health status carefully/correctly to avoid relying on stale data, for example, we can try to access Talos API directly in the controller in addition to checking MachineStatus.

Can only do this for the applies which do not require a reboot.

Further context if needed: https://docs.google.com/document/d/1zFXi8Vut8-qmBoWVl9ZM98nVDZE4F3dj9uamgEtKE6E/edit?tab=t.zr0wf2vhcg5

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    Status

    To Do

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions