Closed BenTheElder closed 1 week ago
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: BenTheElder
The full list of commands accepted by this bot can be found here.
The pull request process is described here
/lgtm
/override pull-kind-e2e-kubernetes-1-29
pod scheduling timeout in prow is not great but also not relevant to this PR
@BenTheElder: Overrode contexts on behalf of BenTheElder: pull-kind-e2e-kubernetes-1-29
This change isn't working yet.
Looks like Fedora 39 works. It's still supported but 40 would be more current.
It also looks like the downloads are slow and we should switch back to downloading from a mirror.
Just capturing failure message here for convenience and in case we lose the job logs.
Pod "pfpod2" is invalid: [spec.containers[0].resources.requests: Invalid value: "600m": must be less than or equal to cpu limit of 500m, spec.containers[0].resources.requests: Invalid value: "600Gi": must be less than or equal to ephemeral-storage limit of 500Gi, spec.containers[0].resources.requests: Invalid value: "600Mi": must be less than or equal to memory limit of 500Mi]
So that's a bug in the upstream tests and we should file an issue there, it's defintiely not related to this PR which only affects some of the github actions.
/override pull-kind-e2e-kubernetes-1-26 pull-kind-e2e-kubernetes-1-27
(prow e2e not relevant for this change)
@BenTheElder: Overrode contexts on behalf of BenTheElder: pull-kind-e2e-kubernetes-1-26, pull-kind-e2e-kubernetes-1-27
actions are passing, I lost context and I can't see how to pull up older runs easily on github for comparison, so I'm not sure if the download is faster but it seems sufficiently fast.
/hold-cancel
actions are passing, I lost context and I can't see how to pull up older runs easily on github for comparison, so I'm not sure if the download is faster but it seems sufficiently fast.
Looking under the Actions tab, I took a look at this and past runs.
The current VM/CGroupv2
runs took anywhere between 5:37 and 5.51 for the "Boot Fedora" step.
The same jobs with the "e2e-k8s.sh: support --ginkgo.label-filter" PR took between 4:35 and 5:43.
So very slightly more time with the newer Fedora, but that image is probably very slightly larger than the older one anyway. That seems reasonable to me.
/lgtm
@BenTheElder: The following tests failed, say /retest
to rerun all failed tests or /retest-required
to rerun all mandatory failed tests:
Test name | Commit | Details | Required | Rerun command |
---|---|---|---|---|
pull-kind-conformance-parallel-ipv6 | 0ff901d99664e973b1f6d7ac0f68a70e1abea8e6 | link | true | /test pull-kind-conformance-parallel-ipv6 |
pull-kind-e2e-kubernetes-1-29 | 0ff901d99664e973b1f6d7ac0f68a70e1abea8e6 | link | true | /test pull-kind-e2e-kubernetes-1-29 |
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 cancel
~stop hardcoding exact download~
xref https://github.com/kubernetes-sigs/kind/pull/3653
/hold
TODO: figure out why fedora 40 isn't working