Container Runtime (e.g. Docker) Isn't Started

Problem

• The pf9-kube service is failing to connect to the the container runtime daemon (e.g. Docker) with a failure similar to below.

var/log/pf9/kube/kube.log

Starting kubelet
Ensuring container 'proxy' is destroyed
docker: Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?.
See 'docker run --help'.
Waiting for "docker -H unix:///var/run/docker.sock          run --name proxy          --detach=true        --net=host        --privileged        --volume /etc/pf9/kube.d/kubeconfigs/kube-proxy.yaml:/etc/kubernetes/pf9/kube-proxy/kube-proxy.yaml gcr.io/google_containers/hyperkube:v1.12.7 /hyperkube proxy        --kubeconfig=/etc/kubernetes/pf9/kube-proxy/kube-proxy.yaml        --v=2        --cluster-cidr 10.20.0.0/16" to evaluate to true ...
docker: Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?.

• The container runtime (docker – in this case) service shows as "inactive (dead)".

None

$ sudo systemctl status docker
● docker.service - Docker Application Container Engine
Loaded: loaded (/usr/lib/systemd/system/docker.service; disabled; vendor preset: disabled)
Active: inactive (dead)
    Docs: https://docs.docker.com

Environment

• Platform9 Managed Kubernetes - All Versions

• Container Runtime

  • Docker

Cause

The container runtime is not being managed by the pf9-kube service.

Resolution

1. In /etc/pf9/kube_override.env, dictate that the PF9_MANAGED_DOCKER variable be exported as " true".

None

echo export PF9_MANAGED_DOCKER="true" > /etc/pf9/kube_override.env

Additional Information

auto$