feat: transform imagePullPolicy when using local cluster

[lucasrod16]

Fixes: #6992

Description

Introduces a ReplaceImagePullPolicy function to update the imagePullPolicy field in Kubernetes manifests to Never when running on a local cluster. This prevents deployment errors that occur when imagePullPolicy is set to Always, as detailed in #6992.

User facing changes

Files to reproduce:

• Dockerfile

  FROM nginx:alpine
  

• pod.yaml

  apiVersion: v1
  kind: Pod
  metadata:
    name: nginx
  spec:
    containers:
    - name: nginx
      image: nginx:alpine
      imagePullPolicy: Always
  

• skaffold.yaml

  apiVersion: skaffold/v4beta11
  kind: Config
  metadata:
    name: nginx
  build:
    artifacts:
      - image: nginx
        context: .
        docker:
          dockerfile: Dockerfile
  manifests:
    rawYaml:
     - ./pod.yaml
  

Before:

After:

[google-cla[bot]]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[lucasrod16]

Hi @ericzzzzzzz,

I noticed that this pull request has been pending for a while. Is there anything needed from me? Please let me know if there are any updates required.

Thanks!

[ericzzzzzzz]

Hi, @lucasrod16 Thank you for the pr!

I think we should only do this transformation when using skaffold dev, skaffold run, skaffold debug as we know that the transformed manifests will be applied to the connected local cluster, but for skaffold render command, we're not sure how the manifests will be used, the rendered results may be used in other places even without skaffold.

Also, this pr does the requested change when kubectl deployer is used, it needs to do the similar thing when helm deployer is used.

[lucasrod16]

Hi, @lucasrod16 Thank you for the pr!

I think we should only do this transformation when using skaffold dev, skaffold run, skaffold debug as we know that the transformed manifests will be applied to the connected local cluster, but for skaffold render command, we're not sure how the manifests will be used, the rendered results may be used in other places even without skaffold.

Also, this pr does the requested change when kubectl deployer is used, it needs to do the similar thing when helm deployer is used.

@ericzzzzzzz That makes sense, thanks for the feedback!

I moved the transform logic from BaseTransform to kubectl Deploy and verified it works locally with dev, run, and debug.

I will work on adding the transform logic to the helm deployer as well

[mepi262]

@lucasrod16 The reviewer has commented to your pull request. How about check it?

[lucasrod16]

Hey @alphanota, thanks for the PR review! I've updated the PR description to ensure the related issue isn't closed when this PR is merged and have addressed all your feedback. I currently don't have the bandwidth to implement the helm functionality, and I appreciate your understanding. Thanks again for the feedback!

[plumpy]

Given the issue reported in #9687, this appears to cause a breaking change. I'm going to go ahead and revert this PR.

I suspect what needs to happen is you only want to replace the pull policy on images that were actually built by Skaffold as part of a skaffold dev command. But it might need more discussion.