Fix TestHostNetworkEndpointPublishingStrategy

[frobware]

The API change in openshift/api#1097 does not mark the set of configurable ports on type HostNetworkStrategy as optional. It's not clear whether that was deliberate and/or intended but without specifying these ports the e2e test TestHostNetworkEndpointPublishingStrategy() will always fail.

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: frobware

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

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [frobware]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[frobware]

error: error creating buildah builder: initializing source docker://image-registry.openshift-image-registry.svc:5000/ci-op-c2k6zvk7/pipeline@sha256:30ea73ec25783d6c04b6648a2e2ae60022f1c5f71bd296e2c4e84b9de840841a: pinging container registry image-registry.openshift-image-registry.svc:5000: Get "https://image-registry.openshift-image-registry.svc:5000/v2/": dial tcp: i/o timeout

/retest

[frobware]

cc @arjunrn

[arjunrn]

@frobware I'm not sure I understand the cause of the failures. This repo does not contain the changes I made in openshift/api. So the CRD being created in the test cluster will not have the new fields. Secondly the ports are optional implicitly IIUC. Only fields which are listed in required are actually required.

cc @alebedev87

[arjunrn]

Anyway, there is some defaulting for the values for the controller when it's not present(due to version upgrades) in my PR #694. If that's the cause of the issue then the defaulting should help.

[frobware]

Anyway, there is some defaulting for the values for the controller when it's not present(due to version upgrades) in my PR #694. If that's the cause of the issue then the defaulting should help.

Without +optional and ,omitempty we don't get any defaults.

[frobware]

@frobware I'm not sure I understand the cause of the failures. This repo does not contain the changes I made in openshift/api.

This is a good point. I currently have stacked API changes. I was trying to understand why my PR #735 was failing in CI. I have bumped to include my changes, but that will pick up yours too.

[Miciah]

spec.endpointPublishingStrategy.hostNetwork is an optional pointer-type field, so I wouldn't expect omitting the field entirely to be a problem. Do you have an error message from a CI failure with a Create or Update call that fails because of the missing values?

[Miciah]

TestHostNetworkEndpointPublishingStrategy did not fail in the last e2e-aws-operator run.

/test e2e-aws-operator

[frobware]

2022-04-12T16:46:23.561+0100	INFO	operator.ingress_controller	controller/controller.go:114	reconciling	{"request": "openshift-ingress-operator/host"}
2022-04-12T16:46:24.259+0100	ERROR	operator.init.controller.ingress_controller	controller/controller.go:266	Reconciler error	{"name": "host", "namespace": "openshift-ingress-operator", "error": "failed to admit ingresscontroller: failed to update status: IngressController.operator.openshift.io \"host\" is invalid: [status.endpointPublishingStrategy.hostNetwork.httpsPort: Invalid value: 0: status.endpointPublishingStrategy.hostNetwork.httpsPort in body should be greater than or equal to 1, status.endpointPublishingStrategy.hostNetwork.statsPort: Invalid value: 0: status.endpointPublishingStrategy.hostNetwork.statsPort in body should be greater than or equal to 1, status.endpointPublishingStrategy.hostNetwork.httpPort: Invalid value: 0: status.endpointPublishingStrategy.hostNetwork.httpPort in body should be greater than or equal to 1]"}

This is the error I see when creating the following:

apiVersion: operator.openshift.io/v1
kind: IngressController
metadata:
  creationTimestamp: "2022-04-12T15:14:20Z"
  finalizers:
  - ingresscontroller.operator.openshift.io/finalizer-ingresscontroller
  generation: 1
  name: host
  namespace: openshift-ingress-operator
  resourceVersion: "105051"
  uid: 93e19982-6366-4345-a1fe-d3cadbd8d712
spec:
  clientTLS:
    clientCA:
      name: ""
    clientCertificatePolicy: ""
  domain: foo.host.amcdermo-2022-04-12-1246.devcluster.openshift.com
  endpointPublishingStrategy:
    type: HostNetwork
    hostNetwork:

which is, as I understand things, the reason why the test fails because we don't currently specify any values for the 3 ports.

[frobware]

@arjunrn In dff1967#diff-0c9271ea6402ecb6233e9e0a629cd72156e12bb7a2c41ef99f0d0265394d7ec4R391 we add default values if they are not specified. Why don't we allow 0 in the API and get the defaults from there?

[openshift-ci]

@frobware: 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
ci/prow/e2e-aws-single-node	fdb19f1	link	false	/test e2e-aws-single-node
ci/prow/e2e-upgrade	fdb19f1	link	true	/test e2e-upgrade

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

[frobware]

Obsoleted by openshift/api#1175.