[BUG] kubectl - memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused

[jonbcampos-alto]

Affected builder image

name: 'gcr.io/google.com/cloudsdktool/cloud-sdk' entrypoint: 'kubectl'

call:

args:
- "rollout"
- "restart"
- "deployment"
- "-l"
 - "app=myLabel"

Expected Behavior

works (?). runs the rollout

Actual Behavior

I get an error:

Already have image (with digest): gcr.io/google.com/cloudsdktool/cloud-sdk
E0404 20:29:08.702878       1 memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused
E0404 20:29:08.704245       1 memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused
E0404 20:29:08.705268       1 memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused
E0404 20:29:08.707955       1 memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused
E0404 20:29:08.709899       1 memcache.go:265] couldn't get current server API group list: Get "http://localhost:8080/api?timeout=32s": dial tcp 127.0.0.1:8080: connect: connection refused
The connection to the server localhost:8080 was refused - did you specify the right host or port?

Steps to Reproduce the Problem

1. create a cloudbuild file
2. set the environment variables
3. run the following (ish) command

- name: "gcr.io/google.com/cloudsdktool/cloud-sdk"
  entrypoint: 'kubectl'
  args:
  - "rollout"
  - "restart"
  - "deployment"
  - "-l"
  - "app=mylabel"
  env:
  - "CLOUDSDK_COMPUTE_ZONE=${_ZONE}"
  - "CLOUDSDK_CONTAINER_CLUSTER=${_CLUSTER}"
  - "CLOUDSDK_CORE_PROJECT=${_PROJECT}"

[kifeo]

for me this was due to a bad env : .kube/config instead of ~/.kube/config in .bashrc

[caddac]

It is my understanding that the kubectl entrypoint runs the kubectl binary and you need to provide a valid kubeconfig to connect to your cluster. I can reproduce this behavior if I don't have a step which gets credentials first. I added a step like

- id: 'get context'
  name: 'gcr.io/google.com/cloudsdktool/cloud-sdk'
  entrypoint: 'gcloud'
  args: [ 'container', 'clusters', 'get-credentials', 'my-cluster', '--region', 'northamerica-northeast1' ]

and my subsequent apply using the kubectl entrypoint worked just fine.

[haroonc]

Thank @caddac for sharing detailed example.