Hi @tnqn. Thanks for your PR.

I'm waiting for a containerd member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

We saw IP addresses were leaked when using Containerd 1.4.4 and Antrea as the CNI, the symptom is that containerd cannot create sandbox container successfully after running a while, failed because of the error:

msg="Failed to destroy network for sandbox \"27902e644b711ed94ac9699f3f588e98a5d28d1dd6a8a25c2f6e829f2df49da3\"" error="netplugin failed with no error message: context canceled"
msg="RunPodSandbox for &PodSandboxMetadata{Name:nginx-rs-new-7d95d89897-nqkcl,Uid:1f5c3c02-b06e-4985-9a65-46f654b0fc57,Namespace:default,Attempt:0,} failed, error" error="failed to start sandbox container task \"27902e644b711ed94ac9699f3f588e98a5d28d1dd6a8a25c2f6e829f2df49da3\": context canceled: unknown"

The reason that it failed to start sandbox container task may be opencontainers/runc#2865 and may have been fixed by opencontainers/runc#2871.
But on CNI side it didn't receive any CNI DEL call when containerd logged "netplugin failed with no error". I think it's because the context already expired so it failed immediately. #5438 had the same symptom with Weave as the CNI.
Other cleanup operations use a new context to avoid this issue, I think it should apply to CNI cleanup as well.

Build succeeded.

• containerd-build-arm64 : SUCCESS in 5m 25s (non-voting)
• containerd-test-arm64 : SUCCESS in 5m 31s (non-voting)
• containerd-integration-test-arm64 : FAILURE in 26m 54s (non-voting)

/ok-to-test

@fuweid let's make sure k8s CI jobs go green as well.