Helm release errors out on filesystem write because of labeling issue

[binarymason]

Expected behavior

skaffold should add labels to pods that exist from a helm release.

Actual behavior

First skaffold is warning because it is trying to add label to pod that does NOT exist.

WARN[0002] error adding label to runtime object: patching resource default/"elastic-dansn-test": pods "elastic-dansn-test" not found

Then, filesystem writes will cause helm release to crash. I get this error:

Error: UPGRADE FAILED: failed to replace object: Service "elasticsearch-master" is invalid: spec.clusterIP: Invalid value: "": field is immutable

Information

• Skaffold version: v1.10.1
• Operating system: Ubuntu 20.04 LTS
• Contents of skaffold.yaml:

---

apiVersion: skaffold/v2beta4
kind: Config
deploy:
  helm:
    releases:
      - name: elastic
        chartPath: elastic/elasticsearch
        remote: true

• Screenshot:
  

• Possibly related: Remote helm chart shouldn't be upgraded on code changes #2245

Steps to reproduce the behavior

1. Add the Elastic Helm charts repo: helm repo add elastic https://helm.elastic.co
2. Run skaffold dev --port-forward
3. Write to skaffold.yml with no changes (vim :w).

[briandealwis]

Thanks for the report! The upgrade error should be fixed by #3274, which will be in the next release.

Duplicate of #2245