kubectl apply doesn't update nodeSelector

[adampl]

BUG REPORT

Kubernetes version:
Client Version: version.Info{Major:"1", Minor:"11", GitVersion:"v1.11.2", GitCommit:"bb9ffb1654d4a729bb4cec18ff088eacc153c239", GitTreeState:"clean", BuildDate:"2018-08-07T23:17:28Z", GoVersion:"go1.10.3", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"11", GitVersion:"v1.11.1", GitCommit:"b1b29978270dc22fecc592ac55d903350454310a", GitTreeState:"clean", BuildDate:"2018-07-17T18:43:26Z", GoVersion:"go1.10.3", Compiler:"gc", Platform:"linux/amd64"}

Environment:

• Hardware configuration: bare metal x86_64
• OS: Ubuntu 16.04.5 LTS
• Kernel: 4.4.0-131-generic
• Install tools: Rancher 2.0

What happened:
When I apply a modified YAML file for an existing Deployment, all changes are received by the cluster except nodeSelector.

What you expected to happen:
nodeSelector should be updated.

How to reproduce it:

1. Write a deployment file with some node selector in the pod spec template
2. Apply the deployment: kubectl apply -f path/to/file.yaml (should create the deployment)
3. Modify the nodeSelector (eg. add new entry) and another parameter (eg. Replicas)
4. Again apply the deployment: kubectl apply -f path/to/file.yaml
5. Observe that other changes are reflected, but nodeSelector stays unchanged

[seans3]

/kind bug
/sig cli
/area kubectl
/priority P2

[adampl]

I think it's fixed in 1.12 or it was just a glitch, as now I'm unable to reproduce it.

Current version:
Client Version: version.Info{Major:"1", Minor:"12+", GitVersion:"v1.12.0-rc.1", GitCommit:"3e4aee86dfaf933f03e052859c0a1f52704d4fef", GitTreeState:"clean", BuildDate:"2018-09-18T21:08:06Z", GoVersion:"go1.10.3", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"11", GitVersion:"v1.11.2", GitCommit:"bb9ffb1654d4a729bb4cec18ff088eacc153c239", GitTreeState:"clean", BuildDate:"2018-08-07T23:08:19Z", GoVersion:"go1.10.3", Compiler:"gc", Platform:"linux/amd64"}

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[seans3]

/remove-lifecycle rotten

If we're not able to reproduce this in current versions of kubectl, we'll close this.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[seans3]

/remove-lifecycle stale

[pswica]

@seans3
I think @adampl is correct in his earlier statement. I tested just now and could not reproduce.

Tried a few scenerios including:

1. without and then with nodeSelector
2. changing nodeSelector

I think we can close.

[seans3]

@pswica Thanks for the help--closing.

/close

[k8s-ci-robot]

@seans3: Closing this issue.

In response to this:

@pswica Thanks for the help--closing.

/close

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.

[dsymonds]

Hey, we've been observing this on GKE, with 1.13.7-gke.8 for master/nodes and v1.13.7 kubectl. I added a nodeSelector, which got applied, but then removed it from the YAML later on and any future kubectl apply operations (that could successfully change other things) didn't touch nodeSelector.