failed pulling image: Get https://k8s.gcr.io/v2/: Client.Timeout exceeded while awaiting headers

[WhoisZihan]

BUG REPOR

I know this is the same as #2018, but with latest version it still doesn't work and I can only start with --bootstrapper=localkube, other bootstrappers never work.

Minikube version : v0.28.2

• OS : Ubuntu 16.04.5 LTS (Xenial Xerus)
• VM Driver : kvm2 (But virtualbox does not work either)
• ISO version : minikube-v0.28.1.iso, boot2docker.iso
• Install tools:
  • kubectl version: v1.11.2
  • kubeadm version: v1.10.0
  • kubelet version: v1.10.0
• Others:
  The above can be generated in one go with the following commands (can be copied and pasted directly into your terminal):

OS:
NAME="Ubuntu"
VERSION="16.04.5 LTS (Xenial Xerus)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 16.04.5 LTS"
VERSION_ID="16.04"
HOME_URL="http://www.ubuntu.com/"
SUPPORT_URL="http://help.ubuntu.com/"
BUG_REPORT_URL="http://bugs.launchpad.net/ubuntu/"
VERSION_CODENAME=xenial
UBUNTU_CODENAME=xenial

VM driver:
    "DriverName": "kvm2",

ISO version
        "Boot2DockerURL": "file:///home/fma/.minikube/cache/iso/minikube-v0.28.1.iso",
        "ISO": "/home/fma/.minikube/machines/minikube/boot2docker.iso",

What happened:
The command minikube start --vm-driver kvm2 --bootstrapper=kubeadm will get stuck at Starting cluster components.... By checking minikube logs, I find it gets refused when trying to connect to 127.0.0.1:8443: getsockopt: connection refused

What you expected to happen:
It starts normally.

How to reproduce it (as minimally and precisely as possible):

1. Download minikube v0.28.2 and kubelet v1.11.2
2. type minikube start --vm-driver kvm2 --bootstrapper=kubeadm
3. Waiting for it to get stuck

Output of minikube logs (if applicable):
Here is part of it

Sep 05 12:24:15 minikube kubelet[2686]: E0905 12:24:15.468704 2686 pod_workers.go:186] Error syncing pod 31cf0ccbee286239d451edb6fb511513 ("kube-scheduler-minikube_kube-system(31cf0ccbee286239d451edb6fb511513)"), skipping: failed to "CreatePodSandbox" for "kube-scheduler-minikube_kube-system(31cf0ccbee286239d451edb6fb511513)" with CreatePodSandboxError: "CreatePodSandbox for pod "kube-scheduler-minikube_kube-system(31cf0ccbee286239d451edb6fb511513)" failed: rpc error: code = Unknown desc = failed pulling image "k8s.gcr.io/pause-amd64:3.1": Error response from daemon: Get https://k8s.gcr.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)"

Sep 05 12:24:15 minikube kubelet[2686]: E0905 12:24:15.608771 2686 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/kubelet.go:451: Failed to list *v1.Service: Get https://localhost:8443/api/v1/services?limit=500&resourceVersion=0: dial tcp 127.0.0.1:8443: getsockopt: connection refused

Anything else do we need to know:

1. I can start with --bootstrapper=localkube but not with kubeadm, will it be the issue of RBAC?
2. I can do ping k8s.io, but I can NOT ping k8s.gcr.io, will this cause start failure too?
3. I also tried specifying --kubernetes-version=v1.9.4, but not working either.

[tstromberg]

Do you mind attaching a full log file? There seems to be two main issues here:

• Unable to deploy kube-scheduler because it can't access k8s.gcr.io.
• The apiserver isn't running (:8443 connect refused)

The second can be many things. The first is indicative of a connectivity failure to GCR: is this minikube instance running within China, by any chance?

We should find a way to install minikube so that it doesn't require external connectivity.

[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