made sure that manager is started before staleController

[DarkInfernus]

Bug fix for the issue #6152

[antoninbas]

@DarkInfernus do you have logs from before and after the change, so we can confirm that the fix is valid?

[DarkInfernus]

Hi @antoninbas actually this is my first ever open source contribution, my pr is failing multiple checks, actually i did not check for the previous logs and after logs. I just looked at the code and as per my knowledge of go lang i think this error is happening because we are starting the go routine for running stalecontroller before starting the manager. So i thought it to be intuitive to start the go routine later.

[DarkInfernus]

But final call should be made after checking the before logs and after logs.For that I would need to set up the project locally but unfortunately I don't know how to do that.
Can you guide me to the correct resource for doing that?
Also can you tell me when this error is triggered so that i can stimulate the situation locally after setting up the project.

[antoninbas]

@DarkInfernus The only way to fix this issue is to first reproduce it locally, understand the code and make the necessary changes, then confirm the issue has been resolved by testing the updated code.
The issue can be reproduced by following the steps in the multi-cluster documentation: https://github.com/antrea-io/antrea/blob/main/docs/multicluster/quick-start.md
I have to warn you that while the fix itself is probably very simple, coming up with the right fix and validating it may not be trivial and may be a bit time-consuming.

[luolanzone]

Hi, I don't think this change will fix the issue, please check my comment on the issue #6152 (comment).
Even we started this after the Controller Manager, it doesn't guarantee that caches are ready before it.

[DarkInfernus]

Ok , i will make the required changes in the code and raise another pr.