Can you please rebase?

/ok-to-test

@wojtek-t is there a way to run any of the kubemark scale tests against a PR to check resource consumption? I'd like to compare CPU use on the apiserver with this enabled

@liggitt are these known test failures? they do not seem related to the change

is it worth it to bump this feature from alpha to beta or better?

is it worth it to bump this feature from alpha to beta or better?

1.8 is the first release to include it. I think it should stay as-is for 1.8, and be bumped up and/or enabled by default at the beginning of the 1.9 cycle for soak time.

/retest

/test pull-kubernetes-e2e-kops-aws

I'm not sure if this is an official policy, but I think features should be in beta stage before they're enabled by default. If the goal is just to enable the feature in the E2E environment, use the --feature-gate flag:

I’d actually advocate switching the feature to beta and enabling by default now to get soak time during 1.9

@liggitt updated to beta feature

/retest

/retest

This needs to be rebased without the merge commit

@smarterclayton done

@ilackarms: The following test failed, say /retest to rerun them all:

Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR.

/hold
still seeing log streaming issues with this enabled in openshift
need to triage before enabling

/approve

Giving the tag for when it is ready to go.

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ilackarms, lavalamp, liggitt, tallclair

Associated issue: #46966

The full list of commands accepted by this bot can be found here.

@LiGgit what needs to be done to move this forward?

still seeing log streaming issues with this enabled in openshift
need to triage before enabling

where/how did you see these issues? where would one start to investigate this?

@liggitt Friendly ping, is there a way for someone not experienced in k8s internals to help here?

the test we saw failing was streaming logs from a live pod: https://openshift-gce-devel.appspot.com/pr/18124

all five runs failed the same way: https://openshift-gce-devel.appspot.com/build/origin-ci-test/pr-logs/pull/18124/test_pull_request_origin_extended_conformance_install/5735/

/tmp/openshift/build-rpm-release/rpm/BUILD/origin-3.9.0/_output/local/go/src/github.com/openshift/origin/test/extended/deployments/deployments.go:783
Expected
    <string>: --> pre: Running hook pod ...
    --> pre: Retrying hook pod (retry #1)
to contain substring
    <string>: pre hook logs
/tmp/openshift/build-rpm-release/rpm/BUILD/origin-3.9.0/_output/local/go/src/github.com/openshift/origin/test/extended/deployments/deployments.go:794

I don't know if we have tests in kube CI for log streaming... the ones I see appear to timeout and restart, so would eventually pick up all logs from the pod when the pod exited.

To recreate/make progress, I think I would look to craft a streaming log test that did the following:

• craft a pod that blocks until it gets a signal, then produces multiple lines of predictable output, then pauses, then exits
• start the pod
• wait until it is running
• start a log streamer (like kubelet_test does: podClient.GetLogs(podName, &v1.PodLogOptions{}).Stream())
• give the pod the signal to continue
• verify the expected output is received by the stream prior to pod exit

that would ensure the content is actually getting flushed through the pod->CRI->kubelet->apiserver->client pipeline, even with gzip on.

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

Before enabled default 'true' for APIResponseCompression feature flag --- need to fix this:
#54205

/remove-lifecycle rotten

/lifecycle frozen

closing in favor of #77449

/close

@liggitt: Closed this PR.