openstack: handle when metadata service returns empty data

[yasminvalim]

The fix removes the marshalli, now the code return raw bytes directly instead of marshalling parsed configs.

[gemini-code-assist]

Code Review

This pull request correctly addresses an issue where empty metadata from the OpenStack service was not handled, leading to errors. The fix is straightforward and effective, adding a check for errors.ErrEmpty. The inclusion of a new test case to cover this scenario is a great addition. I have one suggestion to further improve the robustness of the error handling.

[prestist]

So I dont think this is the correct way to handle our issue unfortunately.

We will need release notes under fixes or bug fixes as we are fixing a bug in an older release.

Regarding the contents of the changes.

An easy way to remove the whole unmarshaling thing and pull the thread of what needs to change would be to remove the "encoding/json" import. That should highlight the unmarshaling code.

fetchConfigFromMetadataServiceIPv4Only and fetchConfigParallel both need changes.

The former needs to return unaltered data and the latter needs a signature change "return []byte" and also ensure unaltered data.

[yasminvalim]

@prestist Thanks for all information. I update the code to return raw bytes from fetchConfigParallel and fetchConfigFromMetadataServiceIPv4Only with no marshalling. Now we are not using FetchConfigDualStack because it returns a parsed types.Config anymore. I guess this is enough to fix the bug, but I need to test it manually and run the kola tests locally to be sure.

Can you take a look and see if this new changes makes sense?

[prestist]

@prestist Thanks for all information. I update the code to return now return raw bytes from fetchConfigParallel and fetchConfigFromMetadataServiceIPv4Only with no marshalling. Now we are not using FetchConfigDualStack because it returns a parsed types.Config anymore. I guess this is enough to fix the bug, but I need to test it manually and run the kola tests locally to be sure.

Can you take a look and see if this new changes makes sense?

Sure taking a look now :)

[prestist]

This is about what I was expecting, the main thing is to ensure dual stack still works (it should not really a big diffrence here) and then verify it fixed the kola test that was failing.

Overall LGTM, thank you for fixing this so quick!

Once you have tested it I will approve :)

[prestist]

@marmijo and @yasminvalim verified the efficacy of these changes and we can land them.