Make configure_controller lifecycle transition strict (no multi-hop)

[vedh1234]

Fix for issue #2916
As discussed in #2916 and #2414, configure_controller() previously performed a multi-hop transition: if a controller was in Inactive state, it would auto-cleanup before configuring.
For now cleanup sequence in unload_controller was kept as-is. Altering this would strictly break backward compatibility for service clients (e.g., spawner, orchestrator nodes).

Changes:

• configure_controller(): now only accepts Unconfigured state; any other state returns ERROR
• Removed the implicit auto-cleanup of Inactive controllers
• Updated tests to reflect the strict behaviour

[github-actions]

This PR is stale because it has been open for 45 days with no activity. Please tag a maintainer for help on completing this PR, or close it if you think it has become obsolete.

[christophfroehlich]

@vedh1234 could you follow-up on the comments please?

[vedh1234]

Sorry for the delay.
I have updated the required changes also while looking into code found that two tests (inactive_controller_cannot_be_cleaned_up and inactive_controller_cannot_be_configured) were verifying the same thing so I have merged them into single configure_controller_requires_unconfigured_state test. I have rebased it as well.

[saikishor]

@vedh1234 Can you recheck the tests are failing

[vedh1234]

Hi @saikishor!
I have pushed a fix for TestControllerManagerSrvs.configure_controller_srv, which wasn't updated when configure_controller became strict. Please approve the workflow run.
Thanks!

[christophfroehlich]

Changes LGTM, but the tests are failing

[vedh1234]

I'm looking into it, on my local run, all the tests are passing (except the hardware related) but for some reason here they are failing. I'll check the CI logs.