New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
deflake grace period for pod deletion in node e2e #89379
deflake grace period for pod deletion in node e2e #89379
Conversation
will discuss in sig-node 3/24 |
/milestone v1.18 |
/assign @dchen1107 |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: derekwaynecarr The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
8c1956c
to
02e4ce6
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I missed the meeting, but some small readability nits. Thanks for doing this - in favor of any change that helps make the tests more stable.
@@ -55,9 +55,9 @@ var _ = SIGDescribe("Pods Extended", func() { | |||
/* | |||
Release : v1.15 | |||
Testname: Pods, delete grace period | |||
Description: Create a pod, make sure it is running. Using the http client send a 'delete' with gracePeriodSeconds=30. Pod SHOULD get deleted within 30 seconds. | |||
Description: Create a pod, make sure it is running. Using the http client send a 'delete' with gracePeriodSeconds=30. Pod SHOULD get terminated within gracePeriodSeconds and removed from API server within a window. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nit: Worth adding a ballpark range for the window value? I.e. is it minutes, 10s of minutes, etc...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
the window is not specific right now, 30s is too tight. 2x30s was still not perfect based on @liggitt past attempts, but trying 3x30s to see if we can still ensure the function works while we work out more precise SLA for kubelet deleting the pod from the API server.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Gotcha! Could be helpful to have that info in the commit message or as a comment so someone looking at this in the future doesn't need to return to the pr description.
@@ -117,8 +117,11 @@ var _ = SIGDescribe("Pods Extended", func() { | |||
|
|||
ginkgo.By("verifying the kubelet observed the termination notice") | |||
|
|||
// allow up to 3x grace period (which allows process termination) | |||
// for the kubelet to remove from api. need to follow-up on if this |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nit: Worth adding a brief description, either here or in the commit message, around how we arrived at 3 as the multiplier?
ok, well, at least we didnt flake on this test. would prefer we merge this so we can keep testing core function, and then in sig-node we need to figure out a plan to establish more precise metrics for kubelet reporting status changes back to api server on pod deletion. |
/hold cancel |
/test pull-kubernetes-e2e-gce |
/assign tallclair |
/lgtm Doesn't seem ideal, but I don't have a better solution. Do we have any ideas why it might take so long to terminate? |
What type of PR is this?
/kind flake
What this PR does / why we need it:
Timing window appears too tight in e2e setup for kubelet to have delete recorded with API.
Which issue(s) this PR fixes:
Fixes #85762
Special notes for your reviewer:
Does this PR introduce a user-facing change?: