Wait for Antrea client to be ready before starting watches

[tnqn]

Fixes #1035

[antrea-bot]

Thanks for your PR.
Unit tests and code linters are run automatically every time the PR is updated.
E2e, conformance and network policy tests can only be triggered by a member of the vmware-tanzu organization. Regular contributors to the project should join the org.

The following commands are available:

• /test-e2e: to trigger e2e tests.
• /skip-e2e: to skip e2e tests.
• /test-conformance: to trigger conformance tests.
• /skip-conformance: to skip conformance tests.
• /test-whole-conformance: to trigger all conformance tests on linux.
• /skip-whole-conformance: to skip all conformance tests on linux.
• /test-networkpolicy: to trigger networkpolicy tests.
• /skip-networkpolicy: to skip networkpolicy tests.
• /test-windows-conformance: to trigger windows conformance tests.
• /skip-windows-conformance: to skip windows conformance tests.
• /test-windows-networkpolicy: to trigger windows networkpolicy tests.
• /skip-windows-networkpolicy: to skip windows networkpolicy tests.
• /test-hw-offload: to trigger ovs hardware offload test.
• /skip-hw-offload: to skip ovs hardware offload test.
• /test-all: to trigger all tests (except whole conformance).
• /skip-all: to skip all tests (except whole conformance).

These commands can only be run by members of the vmware-tanzu organization.

[tnqn]

/test-all

[antoninbas]

question about error handling, otherwise LGTM

[antoninbas]

Is this the right way to handle the error here? We will just log one error message and return the error (which will be ignored by the caller). I would argue that this is more of a "fatal" error, that should cause the agent to crash, so that it doesn't go unnoticed. What do you think?

[tnqn]

Thanks for pointing it out, I just took another look, wait.PollImmediateUntil only returns error when stopCh is closed, which is part of process exiting, then perhaps the log should just say it no longer waits for it and stop the goroutine gracefully. We shouldn't log errors in other controllers too.
https://github.com/kubernetes/kubernetes/blob/master/staging/src/k8s.io/client-go/tools/cache/shared_informer.go#L266
https://github.com/kubernetes/kubernetes/blob/master/pkg/controller/daemon/daemon_controller.go#L288

[antoninbas]

LGTM. I think my once concern is that if the ConfigMap doesn't become available for a while, e.g. because of a controller issue, we will not get a single message in the logs about it, unless I am missing something. Maybe a log message with an exponential backoff would be nice.

[tnqn]

/test-all

[tnqn]

LGTM. I think my once concern is that if the ConfigMap doesn't become available for a while, e.g. because of a controller issue, we will not get a single message in the logs about it, unless I am missing something. Maybe a log message with an exponential backoff would be nice.

@antoninbas I understand your concern now, I have added an info log to indicate that the controller is waiting for antrea client ready per 2 seconds, I didn't use exponential backoff for the whole check because it still needs to be aggressive since it's important to receive network policy as early as possible. Please see if it addresses your concern.

[antoninbas]

LGTM