Repository navigation
fix(e2e): make shared supervisor harness work with gateway TLS - #4984
Merged
drew merged 1 commit intoOct 28, 2026
Merged
Conversation
The TLS port-forward readiness probe sent plaintext HTTP, which the gateway's TLS listener resets, so the shared-supervisor e2e tasks failed before any test ran. In TLS mode, check only that the forward accepts TCP. Shared supervisor pool members are TCP-ready before they attach to the gateway, and the gateway refuses sandbox creates until one attaches. Wait for every pool member to attach before handing the gateway to the suites. Signed-off-by: divesh <dgude@nvidia.com>
drew
approved these changes
Oct 28, 2026
Collaborator
|
/ok to test 937abc2 |
drew
enabled auto-merge
October 28, 2026 03:49
|
Label |
Collaborator
Author
|
/ok to test 937abc2 |
Collaborator
Author
|
/ok to test 937abc2 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The shared-supervisor e2e tasks (
e2e:kubernetes:shared-supervisor,-ha,-scaling) fail during harness setup, before any test runs. This fixes the gateway port-forward readiness probe for TLS mode and waits for supervisor pool members to attach before the suites start.Related Issue
No issue yet. Failures fixed:
start_gateway_portforwardprobes with plaintextcurl http://.... WithOPENSHELL_E2E_KUBE_TLS=1the gateway's TLS listener resets the connection (curl exit 56), so the probe never passes and setup fails.sandbox createfails with "the shared supervisor has no capacity for another sandbox instance".Changes
start_gateway_portforward: in TLS mode, check that the forward accepts TCP instead of sending plaintext HTTP. Non-TLS mode is unchanged.wait_for_shared_supervisors, called after the health port-forward: a no-op unless shared pool Pods (app.kubernetes.io/component=supervisor) exist. Otherwise waits up to 180 s for every member to loggateway accepted the supervisor host(the signal the shared-supervisor tests use), then 2 s for the membership record.Testing
bash -n e2e/with-kube-gateway.she2e:kubernetes:shared-supervisor,-ha, and-scalingall passed (conformance 8/8 plus each task's test); without it all three fail at setupChecklist