Update openapi-gen to not rely on vendor

[Jeffwan]

1. Update gen-sdk.sh to use kube-openapi from go modules rather than vendor. Not all code-generator has /cmd/openapi-gen, 1.15 doesn't while it's added back in 1.16. Let's use kube-openapi here.
2. central.maven.org is not resolvable on my end, seems it's http legacy url. Update the jar url and points to repo1.maven.org

[Jeffwan]

@jinchihe Do we need to addgithub.com/kubeflow/common/job_controller/api/v1 as input? I actually have some failure if I add common API github.com/kubeflow/common/job_controller/api/v1 like below. Have you ever seen the problem?

$ ./openapi-gen --input-dirs github.com/kubeflow/common/job_controller/api/v1,github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --output-package github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --go-header-file hack/boilerplate/boilerplate.go.txt

E0519 14:11:58.053066   14182 openapi.go:294] Error when generating: PodInformerSynced, PodInformerSynced k8s.io/client-go/tools/cache.InformerSynced
2020/05/19 14:11:58 OpenAPI code generation error: Failed executing generator: some packages had errors:
cannot generate spec for type k8s.io/client-go/tools/cache.InformerSynced

[gaocegege]

cc @jinchihe

[jinchihe]

Sorry I missed the PR since I was on vacation last month, will take a look asap. Thanks,

[jinchihe]

@jinchihe Do we need to addgithub.com/kubeflow/common/job_controller/api/v1 as input? I actually have some failure if I add common API github.com/kubeflow/common/job_controller/api/v1 like below. Have you ever seen the problem?

$ ./openapi-gen --input-dirs github.com/kubeflow/common/job_controller/api/v1,github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --output-package github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --go-header-file hack/boilerplate/boilerplate.go.txt

E0519 14:11:58.053066   14182 openapi.go:294] Error when generating: PodInformerSynced, PodInformerSynced k8s.io/client-go/tools/cache.InformerSynced
2020/05/19 14:11:58 OpenAPI code generation error: Failed executing generator: some packages had errors:
cannot generate spec for type k8s.io/client-go/tools/cache.InformerSynced

@Jeffwan I do not remember hit the problem. If the problem can be resolved by adding the input , yeah, please go ahead.

[jinchihe]

@jinchihe Do we need to addgithub.com/kubeflow/common/job_controller/api/v1 as input? I actually have some failure if I add common API github.com/kubeflow/common/job_controller/api/v1 like below. Have you ever seen the problem?

$ ./openapi-gen --input-dirs github.com/kubeflow/common/job_controller/api/v1,github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --output-package github.com/kubeflow/pytorch-operator/pkg/apis/pytorch/v1 --go-header-file hack/boilerplate/boilerplate.go.txt

E0519 14:11:58.053066   14182 openapi.go:294] Error when generating: PodInformerSynced, PodInformerSynced k8s.io/client-go/tools/cache.InformerSynced
2020/05/19 14:11:58 OpenAPI code generation error: Failed executing generator: some packages had errors:
cannot generate spec for type k8s.io/client-go/tools/cache.InformerSynced

@Jeffwan I do not remember hit the problem. If the problem can be resolved by adding the input , yeah, please go ahead.

[jinchihe]

Would you mind to rebase to remove this change. I did not realized the PR, and already fixed this last week. Thanks a lot!

[Jeffwan]

Sure. I will do.

[Jeffwan]

@jinchihe Please have another check

[jinchihe]

/lgtm

@Jeffwan Thanks!

[Jeffwan]

/cc @johnugeorge @gaocegege

[johnugeorge]

Sorry for being late
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: johnugeorge

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:

• OWNERS [johnugeorge]

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

[Jeffwan]

/test kubeflow-pytorch-operator-presubmit

[Jeffwan]

Test failure is probably related with disk out of capacity failure.