Kubelet and API server is not running.

[rama kumar chittala]

Hi,

When iam trying to run kubelet service on my Oracle Virtual Box coreos machine the following error is occure.

localhost ~ # systemctl status kubelet

● kubelet.service

   Loaded: loaded (/etc/systemd/system/kubelet.service; enabled; vendor preset: disabled)

   Active: activating (auto-restart) (Result: exit-code) since Tue 2017-02-07 15:00:00 UTC; 321ms ago

  Process: 14759 ExecStart=/usr/lib/coreos/kubelet-wrapper --api-servers=http://10.x.x.x:8080 --network-plugin-dir= --network-plugin= --re

  Process: 14756 ExecStartPre=/usr/bin/mkdir -p /etc/kubernetes/manifests (code=exited, status=0/SUCCESS)

 Main PID: 14759 (code=exited, status=254)

Feb 07 15:00:00 localhost systemd[1]: kubelet.service: Main process exited, code=exited, status=254/n/a

Feb 07 15:00:00 localhost systemd[1]: kubelet.service: Unit entered failed state.

Feb 07 15:00:00 localhost systemd[1]: kubelet.service: Failed with result 'exit-code'.

Error log

localhost ~ # journalctl -fu kubelet

-- Logs begin at Mon 2017-02-06 11:38:45 UTC. --

Feb 07 15:00:21 localhost systemd[1]: kubelet.service: Service hold-off time over, scheduling restart.

Feb 07 15:00:21 localhost systemd[1]: Stopped kubelet.service.

Feb 07 15:00:21 localhost systemd[1]: Starting kubelet.service...

Feb 07 15:00:21 localhost systemd[1]: Started kubelet.service.

Feb 07 15:00:21 localhost kubelet-wrapper[14881]: KUBELET_VERSION environment variable is deprecated, please use KUBELET_IMAGE_TAG instead

Feb 07 15:00:21 localhost kubelet-wrapper[14881]: + exec /usr/bin/rkt run --trust-keys-from-https --volume etc-kubernetes,kind=host,source=/etc/kubernetes,readOnly=false --volume etc-ssl-certs,kind=host,source=/etc/ssl/certs,readOnly=true --volume usr-share-certs,kind=host,source=/usr/share/ca-certificates,readOnly=true --volume var-lib-docker,kind=host,source=/var/lib/docker,readOnly=false --volume var-lib-kubelet,kind=host,source=/var/lib/kubelet,readOnly=false --volume os-release,kind=host,source=/usr/lib/os-release,readOnly=true --volume run,kind=host,source=/run,readOnly=false --mount volume=etc-kubernetes,target=/etc/kubernetes --mount volume=etc-ssl-certs,target=/etc/ssl/certs --mount volume=usr-share-certs,target=/usr/share/ca-certificates --mount volume=var-lib-docker,target=/var/lib/docker --mount volume=var-lib-kubelet,target=/var/lib/kubelet --mount volume=os-release,target=/etc/os-release --mount volume=run,target=/run --stage1-from-dir=stage1-fly.aci quay.io/coreos/hyperkube:v1.2.4_coreos.1 --exec=/kubelet -- --api-servers=http://10.x.x.x:8080 --network-plugin-dir= --network-plugin= --register-schedulable=false --allow-privileged=true --config=/etc/kubernetes/manifests --hostname-override=10.x.x.x--v=10

Feb 07 15:00:21 localhost kubelet-wrapper[14881]: run: discovery failed

Feb 07 15:00:21 localhost systemd[1]: kubelet.service: Main process exited, code=exited, status=254/n/a

Feb 07 15:00:21 localhost systemd[1]: kubelet.service: Unit entered failed state.

Feb 07 15:00:21 localhost systemd[1]: kubelet.service: Failed with result 'exit-code'.

Please help.

[paul...@coreos.com]

Hi Rama,

I'm guessing the important bit is "discovery failed". Are you setting up etcd first, and using the right values when launching kubelet?

What guide are you following for setup?

[bzub]

Feb 07 15:00:21 localhost kubelet-wrapper[14881]: run: discovery failed

This error is actually coming from rkt which failed to download the hyperkube image that contains the kubelet.  That's odd though because I was able to download the same one with rkt.  Try running the following command on your node and see if it fails again.  If so, check that you have proper internet access from that node.

sudo rkt fetch quay.io/coreos/hyperkube:v1.2.4_coreos.1