Support antctl command for packetcapture

[hangyan]

TODO items:

• wait for Support collect logs for failed agents and controller for supportbundle #3659 merged and reuse it's uncompress function

after #3659 , the CodeQL alert should be fixed and UT coverage will also be improved.

[hangyan]

cc @antoninbas Can you help review this? also maybe merge #3659 first

cc @luolanzone

[luolanzone]

Please fix the antctl e2e test failures: https://github.com/antrea-io/antrea/actions/runs/13198383877?pr=6884

[luolanzone]

@hangyan there are lots of failures in github checks, please check and fix them.

[hangyan]

Please fix the antctl e2e test failures: https://github.com/antrea-io/antrea/actions/runs/13198383877?pr=6884

I disable the flag to run packetcapture sub-command in agentMode but controllerMode will still fail. One must be chosen, or this sub-command wont be registered to the rootCmd. However, seems packetcapture doens't make much sense in agentMode or controllerMode(controllerMode also contains outside of cluster case I think, but the e2e will check the command in controller pod).

The tests expect run antctl packetcapture to be successful. the first error comes from packet number should be larger than 0, that's the default value and it will cause error. Even we set with a valid default value, no src/dest endpoint （ i think Traceflow support this?), or packetcapture is not enabled , the command will still fail.

Can we make an exception in the test-case or we need to provide some sane defaults for it to working in the controllerMode?

cc @luolanzone @antoninbas

[antoninbas]

Please rename option to options everywhere for consistency, otherwise it looks good to me

[hangyan]

Please rename option to options everywhere for consistency, otherwise it looks good to me

sure, will do.

I disable the flag to run packetcapture sub-command in agentMode but controllerMode will still fail. One must be chosen, or this sub-command wont be registered to the rootCmd. However, seems packetcapture doens't make much sense in agentMode or controllerMode(controllerMode also contains outside of cluster case I think, but the e2e will check the command in controller pod).

The tests expect run antctl packetcapture to be successful. the first error comes from packet number should be larger than 0, that's the default value and it will cause error. Even we set with a valid default value, no src/dest endpoint （ i think Traceflow support this?), or packetcapture is not enabled , the command will still fail.

Can we make an exception in the test-case or we need to provide some sane defaults for it to working in the controllerMode?

Any thoughts on this? cc @antoninbas