luxas / kubeadm-installer

A kubeadm installer container for CoreOS, Ubuntu, Debian, CentOS and Fedora

kubeadm installer for CoreOS, Ubuntu, Debian, CentOS and Fedora

MAINTAINER HAS SWITCHED !

Use https://github.com/xakraz/kubeadm-installer for the latest update and PRs

How to run install kubeadm

Given docker already is installed (otherwise, run curl get.docker.com | bash), you can install kubeadm easily!

$ docker run -it \
	-v /etc/cni:/rootfs/etc/cni \
	-v /etc/systemd:/rootfs/etc/systemd \
	-v /opt:/rootfs/opt \
	-v /usr/bin:/rootfs/usr/bin \
	luxas/kubeadm-installer ${your_os_here}

${your_os_here} can be coreos, ubuntu, debian, fedora or centos

How to uninstall/revert

$ docker run -it 	\
	-v /etc/cni:/rootfs/etc/cni \
	-v /etc/systemd:/rootfs/etc/systemd \
	-v /opt:/rootfs/opt \
	-v /usr/bin:/rootfs/usr/bin \
	luxas/kubeadm-installer ${your_os_here} uninstall

What's inside?

• kubeadm v1.5.0-alpha.2.421+a6bea3d79b8bba, see releases here: http://kubernetes.io/docs/admin/kubeadm/
• kubernetes v1.4.4
• cni 07a8a28637e97b22eb8dfe710eeae1344f69d16e

License

MIT

kubeadm-installer's Issues

Pre-existing ETCD cluster needed?

Do I have to create an etcd cluster for the bare metal machines I plan on using first for CoreOS? Or does Kubernetes(By that I mean kubeadm) configure that for me because all the minions end up pointing to the master, anyway?

Also, I cannot post issues on the new maintainer's repo because it's still a fork of this project. Not sure how to solve that on Github but if the new maintainer wants to solve issues, he needs his own repo.

@xakraz

Transfer ownership ?

Hello @luxas

Thanks for having given me the opportunity to maintain this tool.

Would you transfer me the repo so that every other forks refer to the me now ?

Thx

Note: I will have to delete my fork 1st before accepting your transfer. So ping me when you are available and ready to do it :)

Instructions are wrong

Should be:

$ docker run -it -v /etc/cni:/rootfs/etc/cni -v /etc/systemd:/rootfs/etc/systemd -v /opt:/rootfs/opt -v /usr/bin:/rootfs/usr/bin luxas/kubeadm-installer ${your_os_here}

kublet.service looking for kubelet.conf file that doesn't exist

Trying to bootstrap a Container Linux (CoreOS) cluster with kubeadm-installer, thanks for all the good work!
My node's kubelet is failing to start looking for a kubelet.conf file that does not exist on the system. I don't see one generated by the installer, is this a change in kubeadm setup? Should I be providing one for the system?
Logs:

Jan 21 14:23:07 core-01-node systemd[1]: kubelet.service: Scheduled restart job, restart counter is at 1.              
Jan 21 14:23:07 core-01-node systemd[1]: Stopped kubelet: The Kubernetes Node Agent.                                   
Jan 21 14:23:07 core-01-node systemd[1]: Started kubelet: The Kubernetes Node Agent.                                   
Jan 21 14:23:07 core-01-node kubelet-wrapper[1936]: KUBELET_VERSION environment variable is deprecated, please use KUBELET_IMAGE_TAG instead                                                                                                   
Jan 21 14:23:07 core-01-node kubelet-wrapper[1936]: + exec /usr/bin/rkt run --volume opt-cni,kind=host,source=/opt/cni --mount volume=opt-cni,target=/opt/cni --volume etc-cni,kind=host,source=/etc/cni --mount volume=etc-cni,target=/etc/cni
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: Flag --require-kubeconfig has been deprecated, You no longer need to use --require-kubeconfig. This will be removed in a future version. Providing --kubeconfig enables API server mode, om
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: I0121 14:23:08.812095    1936 server.go:182] Version: v1.9.2+coreos.0                                                                                                                      
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: I0121 14:23:08.812326    1936 feature_gate.go:220] feature gates: &{{} map[]}                                                                                                              
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: W0121 14:23:08.812621    1936 server.go:280] --require-kubeconfig is deprecated. Set --kubeconfig without using --require-kubeconfig.                                                      
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: W0121 14:23:08.812979    1936 cni.go:171] Unable to update cni config: No networks found in /etc/cni/net.d                                                                                 
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: I0121 14:23:08.835305    1936 plugins.go:101] No cloud provider specified.                                                                                                                 
Jan 21 14:23:08 core-01-node kubelet-wrapper[1936]: Error: failed to run Kubelet: invalid kubeconfig: stat /etc/kubernetes/kubelet.conf: no such file or directory                                                             

kubelet.service file:

  [Unit]
  Description=kubelet: The Kubernetes Node Agent
  Documentation=http://kubernetes.io/docs/

  [Service]
  Environment="KUBELET_VERSION=v1.9.2_coreos.0"
  Environment="KUBELET_IMAGE_TAG=v1.9.2_coreos.0"
  Environment="RKT_RUN_ARGS=--volume opt-cni,kind=host,source=/opt/cni --mount volume=opt-cni,target=/opt/cni --volume etc-cni,kind=host,source=/etc/cni --mount volume=etc-cni,target=/etc/cni"
  ExecStart=/usr/lib/coreos/kubelet-wrapper --kubeconfig=/etc/kubernetes/kubelet.conf     --require-kubeconfig=true     --pod-manifest-path=/etc/kubernetes/manifests     --allow-privileged=true     --network-plugin=cni     --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin     --cluster-dns=10.96.0.10 --cluster-domain=cluster.local
  Restart=always
  StartLimitInterval=0
  RestartSec=10

  [Install]
  WantedBy=multi-user.target

CoreOS v 1662.0.0
I'm pulling the 'latest' tag on the kubeadm-installer

@xakraz

coreos not working

i am trying to run it on CoreOS
DISTRIB_ID="Container Linux by CoreOS"
DISTRIB_RELEASE=1967.6.0

docker run -it -v /etc/cni:/rootfs/etc/cni -v /etc/systemd:/rootfs/etc/systemd -v /opt:/rootfs/opt -v

/usr/bin:/rootfs/usr/bin luxas/kubeadm-installer coreos
Installed kubectl in /opt/bin/kubectl
Ignoring /opt/bin/kubelet, since it seems to exist already or isn't needed
Installed kubeadm in /opt/bin/kubeadm
Installed CNI binaries in /opt/cni
Installed the kubelet.service in /etc/systemd/system/kubelet.service
/entrypoint.sh: line 97: /etc/systemd/network/50-cni.network: No such file or directory
Installed 50-cni.network in /etc/systemd/network/50-cni.network
Done! Now run this in your terminal to enable docker and kubelet:

systemctl daemon-reload
systemctl enable docker kubelet
systemctl restart docker kubelet

localhost ~ # /opt/bin/kubeadm init
/opt/bin/kubeadm: line 1: syntax error near unexpected token <' /opt/bin/kubeadm: line 1: NoSuchKeyThe specified key does not exist.

No such object: kubernetes-release-dev/ci-cross/v1.6.0-alpha.0.2074+a092d8e0f95f52/bin/linux/amd64/kubeadm

'

cat /opt/bin/kubeadm

NoSuchKeyThe specified key does not exist.

No such object: kubernetes-release-dev/ci-cross/v1.6.0-alpha.0.2074+a092d8e0f95f52/bin/linux/amd64/kubeadm

Please help..
What am I doing something wrong?

kubelet-wrapper doesn't start any manifest

The kubeadm-installer is using the kubelet-wrapper from CoreOS and none of the static pod is started from the manifests. I assume it's a kubelet-wrapper issue as I cannot reproduce it on different OS where standard Kubelet bin is used.
Would there be another way to use kubeadm on CoreOS nodes ?

Attached is the kubelet log with --v=6
kubelet.txt

"Unbound variable" error when using "default" action (not providing any on the CLI)

https://github.com/xakraz/kubeadm-installer/blob/master/files/entrypoint.sh#L168

$ docker run --rm -it \
>   --env="http_proxy=http://XXXX:3128" \
>   --env="https_proxy=http://XXXX:3128" \
>   --env="KUBEADM_RELEASE=v1.6.1" \
>   --volume=/etc/cni:/rootfs/etc/cni \
>   --volume=/etc/systemd:/rootfs/etc/systemd \
>   --volume=/opt:/rootfs/opt \
>   --volume=/usr/bin:/rootfs/usr/bin \
>   xakra/kubeadm-installer coreos

Unable to find image 'xakra/kubeadm-installer:latest' locally
latest: Pulling from xakra/kubeadm-installer
2aecc7e1714b: Pull complete 
8f7b96c2ad68: Pull complete 
5918ece632fc: Pull complete 
Digest: sha256:a5c7f579e7c9f612d1a613e7c3d6330a41748af83fd1fdc572b66c6e9e2be389
Status: Downloaded newer image for xakra/kubeadm-installer:latest

[ENV] 'K8S_VERSION' is empty -> Getting latest release version
[ENV] 'K8S_VERSION' = v1.6.6
[ENV] 'KUBEADM_RELEASE' = v1.6.1
[ENV] 'CNI_RELEASE' is empty -> Getting latest release version
[ENV] 'CNI_RELEASE' = v0.5.2
[Kubeadm] Selected version: v1.6.1
[Kubeadm] KUBEADM_RELEASE v1.6.1 is a Release version
[Kubeadm] Download url: https://storage.googleapis.com/kubernetes-release/release
[CNI] Selected version: v0.5.2
[CNI] CNI_RELEASE v0.5.2 is a Release version
[CNI] Download url: https://github.com/containernetworking/cni/releases/download/v0.5.2/cni-amd64-v0.5.2.tgz

/entrypoint.sh: line 168: $2: unbound variable

[coreos] Missing ${ROOTFS}

On Container Linux 1353.1.0 the script fails with:

/entrypoint.sh: line 97: /etc/systemd/network/50-cni.network: No such file or directory
Installed 50-cni.network in /etc/systemd/network/50-cni.network
(...)

core@kovh ~ $ ls -l /etc/systemd/network/
total 0

That's because of a missing ${ROOTFS}/ on the line cat > /etc/systemd/network/50-cni.network.

I can't send PR because the script distributed with the luxas/kubeadm-installer on the Docker Hub doesn't match the script in the current repo.

socat missing for CoreOS

preflight checks exit with:

core@ip-10-0-1-121 ~ $ sudo kubeadm init
Running pre-flight checks
preflight check errors:
	socat not found in system path