Handling of deprecated robot providerIDs seems wrong

[janiskemper]

TL;DR

There is a document on migrating from the CAPH providerIDs to the ones used by this project. However, it seems to violate the CAPI behavior, where it says "providerID must match the provider ID as seen on the node object corresponding to this machine".

Expected behavior

I would expect to have a migration path from the Syself fork to this project. However, due to the apparent violation of the CAPI rules, this project doesn't seem to be usable for CAPH users with robot servers.

I mainly want to create awareness for the problem.

We are considering using the Syself CCM fork to create an upgrade path to this project.

Observed behavior

Due to the mentioned violation, CAPI/CAPH clusters don't start successfully with robot servers when using this CCM.

Minimal working example

No response

Log output

Additional information

No response