production k8s install supplement

Just Supplement doc

this document just a supplement for official documentation , if you are newer with k8s install you should go to official doc for look up more info . usually you can install by accord official doc step by step smoothly.

this doc may help for you at below scene:

• can’t access network get yum repository from google release and get aliyun release is fast

• how install specify version of kubelet and kubeadm

• can’t access network k8s.gcr.io for pull docker image and pull aliyun image is fast

yum repository problem

replace yum repository config by aliyun

$ vim /etc/yum.repos.d/k8s.repo

replace official google release with below

[kubernetes]
name=Kubernetes repo
baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/
gpgcheck=0
gpgkey=https://mirrors.aliyun.com/kubernetes/yum/doc/yum-key.gpg
enabled=1

then install kubeadm and kubelet will at just some minute

$ yum install kubelet kubeadm -y

install specify version of kubelet

if you want install specify version of kubernetes you should download from rpm package find the baseurl at aliyun yum release which in your yum repository config property name baseurl

$ curl https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/Packages/230df4e2037cc1655d03586f837a4ecaf42fdbee7366ee2956e08ad1abd3ab8f-kubelet-1.16.10-0.x86_64.rpm -o kubelet-1.16.10.rpm
$ rpm i kubelet-1.16.10.rpm

this method also apply for kubeadm

gcr image problem

default kubeadm init images are k8s.gcr.io you could print default images use command like below

$ kubeadm config images list
k8s.gcr.io/kube-apiserver:v1.17.6
k8s.gcr.io/kube-controller-manager:v1.17.6
k8s.gcr.io/kube-scheduler:v1.17.6
k8s.gcr.io/kube-proxy:v1.17.6
k8s.gcr.io/pause:3.1
k8s.gcr.io/etcd:3.4.3-0
k8s.gcr.io/coredns:1.6.5

you have to method to init by replace gcr images with aliyun images === retag image by pull aliyun image

$ docker image pull registry.aliyuncs.com/google_containers/kube-apiserver:v1.17.6
$ docker tag registry.aliyuncs.com/google_containers/kube-apiserver:v1.17.6  k8s.gcr.io/kube-apiserver:v1.17.6

retag all images then

$ kubeadm init

config imageRepository by aliyun when init

you can modify config file or pass arg command line, here I choose config file way

$ kubeadm config print init-defaults > kubeadm.conf

edit this file replace

imageRepository: k8s.gcr.io

with

imageRepository: registry.aliyuncs.com/google_containers

and then

$ kubeadm init --config kubadm.conf

network plugin with flannel

install flannel plugin as official doc described, but here i have two notice for you

• notice 1 (official doc mention it also) For flannel to work correctly, you must pass --pod-network-cidr=10.244.0.0/16 to kubeadm init.

• notice 2 pass arg --iface={xxxx} if your network face is not normal(eth[\d]+) name , if won’t, you will found fail message when plugin working

License

this doc is Open Source software released under the Apache 2.0 license.