[BUG] The entrypoint script in an agent node continuously fails at the line `until kubectl uncordon "$HOSTNAME"; do sleep 3; done` as KUBECONFIG not set #1420

What did you do

• How was the cluster created?
  • k3d cluster create --config "k3d.yaml"

My k3d.yaml config:

apiVersion: k3d.io/v1alpha5
kind: Simple
servers: 1
agents: 1
image: ${K_SERVER_IMAGE}
token: ${K_SERVER_TOKEN}
network: ${K_NET_NAME}
metadata:
  name: ${K_DEV_NAME}

env:
  - envVar: KUBECONFIG=/var/lib/rancher/k3s/agent/kubelet.kubeconfig
    nodeFilters:
      - agent:*
volumes:
  - volume: "${K_SERVER_VOLUME_SOURCE_0}:${K_NODE_VOLUME_TARGET}"
    nodeFilters:
      - server:*
  - volume: "${K_AGENT_VOLUME_SOURCE_0}:${K_NODE_VOLUME_TARGET}"
    nodeFilters:
      - agent:*

kubeAPI:
  host: "${K_SERVER_DOMAIN}"
  hostPort: "${K_SERVER_KUBEAPI_PORT}"
  hostIP: "0.0.0.0" # "127.0.0.1"

registries:
  config: |
    mirrors:
      "${K_REGISTRY_DOMAIN}":
        rewrite: {}
        endpoint:
          - http://${K_NET_IP_REGISTRY}:${K_REGISTRY_PORT}

hostAliases: []

options:
  k3d:
    wait: true
    timeout: "60s"
    disableLoadbalancer: true
    disableImageVolume: false
    disableRollback: false
  k3s:
    nodeLabels:
      - label: cloud.google.com/gke-nodepool=agent
        nodeFilters: ["agent:*"]
      - label: cloud.google.com/gke-nodepool=server
        nodeFilters: ["server:*"]

    extraArgs:
      - nodeFilters: ["server:*"]
        arg: "--disable=traefik"
      - nodeFilters: ["server:*"]
        arg: "--disable=servicelb"
      - nodeFilters: ["server:*"]
        arg: "--datastore-endpoint=${K_PG_URL}"

  kubeconfig:
    updateDefaultKubeconfig: false
    switchCurrentContext: false
  runtime:
    labels: []
    ulimits:
      - name: nofile
        soft: 2000000
        hard: 2000000

• What did you do afterwards?
  • Nothing

What did you expect to happen

To see the agent logs be fine, instead it's plagued with:

E0328 11:09:48.838205     424 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
E0328 11:09:48.838726     424 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
E0328 11:09:48.840152     424 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
E0328 11:09:48.840786     424 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?

Which version of k3d

k3d version v5.6.0
k3s version v1.21.7-k3s1 (default)

I'm deploying k3s version 1.29.x

---

The problem is in the entrypoint script for a k3s agent node, at the line:

until kubectl uncordon "$HOSTNAME"; do sleep 3; done

That continuously fails, because $KUBECONFIG is not set, so it defaults to http:/localhost:8080. Not sure if I'm understanding the situation correctly but that seems to be it