K8S之一：集群搭建

前言：K8S組件概念

etcd - A highly available key-value store for shared configuration and service discovery.
flannel - An etcd backed network fabric for containers.
kube-apiserver - Provides the API for Kubernetes orchestration.
kube-controller-manager - Enforces Kubernetes services.
kube-scheduler - Schedules containers on hosts.
kubelet - Processes a container manifest so the containers are launched according to how they are described.
kube-proxy - Provides network proxy services.

第一部分：節(jié)點(diǎn)規(guī)劃及準(zhǔn)備

搭建4臺(tái)服務(wù)器，最小化安裝Centos 7 x64系統(tǒng)，以下為架構(gòu)示意圖而账，對(duì)應(yīng)地址及域名為

d181.contoso.com/172.16.3.181
d182.contoso.com/172.16.3.182
d183.contoso.com/172.16.3.183
d184.contoso.com/172.16.3.184
網(wǎng)關(guān)及DNS服務(wù)器為172.16.3.1

k8s.png

關(guān)閉防火墻

所有節(jié)點(diǎn)上
$ systemctl stop firewalld
$ systemctl disable firewalld

第二部分：安裝master

以下操作在d181節(jié)點(diǎn)

安裝軟件包胰坟，以下命令會(huì)自動(dòng)安裝docker

$ yum -y install etcd kubernetes

配置etcd

$ vim /etc/etcd/etcd.conf
ETCD_NAME="default"
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
ETCD_LISTEN_CLIENT_URLS="http://0.0.0.0:2379"
ETCD_ADVERTISE_CLIENT_URLS="http://localhost:2379"

配置api server

$ vim /etc/kubernetes/apiserver
KUBE_API_ADDRESS="--insecure-bind-address=0.0.0.0"
KUBE_API_PORT="--port=8080"
KUBELET_PORT="--kubelet_port=10250"
KUBE_ETCD_SERVERS="--etcd_servers=http://127.0.0.1:2379"
KUBE_SERVICE_ADDRESSES="--service-cluster-ip-range=10.254.0.0/16"

# 注意去掉了ServiceAccount防止報(bào)錯(cuò)
KUBE_ADMISSION_CONTROL="--admission_control=NamespaceLifecycle,NamespaceExists,LimitRanger,SecurityContextDeny,ResourceQuota"
KUBE_API_ARGS=""

啟動(dòng)服務(wù)

$ for SERVICES in etcd kube-apiserver kube-controller-manager kube-scheduler; do
    systemctl restart $SERVICES
    systemctl enable $SERVICES
    systemctl status $SERVICES 
done

定義flannel網(wǎng)絡(luò)

$ etcdctl mk /atomic.io/network/config '{"Network":"172.27.0.0/16"}'

測試

$ kubectl get nodes

結(jié)果應(yīng)該什么都沒有，因?yàn)檫€未啟動(dòng)節(jié)點(diǎn)服務(wù)

第三部分：安裝nodes

以下操作在其他節(jié)點(diǎn)

安裝軟件包

$ yum -y install flannel kubernetes

配置與etcd連接

vim /etc/sysconfig/flanneld
FLANNEL_ETCD_ENDPOINTS="http://d181.contoso.com:2379"

配置與apiserver連接

vim /etc/kubernetes/config
KUBE_MASTER="--master=http://d181.contoso.com:8080"

d182配置kubelet服務(wù)

vim /etc/kubernetes/kubelet
KUBELET_ADDRESS="--address=0.0.0.0"
KUBELET_PORT="--port=10250"
# change the hostname to this host’s IP address
KUBELET_HOSTNAME="--hostname_override=d182.contoso.com"
KUBELET_API_SERVER="--api_servers=http://d181.contoso.com:8080"
#KUBELET_POD_INFRA_CONTAINER="--pod-infra-container-image=registry.access.redhat.com/rhel7/pod-infrastructure:latest"
KUBELET_ARGS="--pod-infra-container-image=kubernetes/pause --cluster-domain=contoso.com --cluster-dns=172.16.3.1"
上句如果采用本地鏡像有庫則為
KUBELET_ARGS="--pod-infra-container-image=docker.contoso.com:5000/pause-amd64.3.0 --cluster-domain=contoso.com --cluster-dns=172.16.3.1"

d183配置kubelet服務(wù)

vim /etc/kubernetes/kubelet
KUBELET_ADDRESS="--address=0.0.0.0"
KUBELET_PORT="--port=10250"
# change the hostname to this host’s IP address
KUBELET_HOSTNAME="--hostname_override=d183.contoso.com"
KUBELET_API_SERVER="--api_servers=http://d181.contoso.com:8080"
#KUBELET_POD_INFRA_CONTAINER="--pod-infra-container-image=registry.access.redhat.com/rhel7/pod-infrastructure:latest"
KUBELET_ARGS="--pod-infra-container-image=kubernetes/pause --cluster-domain=contoso.com --cluster-dns=172.16.3.1"

d184配置kubelet服務(wù)

vim /etc/kubernetes/kubelet
KUBELET_ADDRESS="--address=0.0.0.0"
KUBELET_PORT="--port=10250"
# change the hostname to this host’s IP address
KUBELET_HOSTNAME="--hostname_override=d184.contoso.com"
KUBELET_API_SERVER="--api_servers=http://d181.contoso.com:8080"
#KUBELET_POD_INFRA_CONTAINER="--pod-infra-container-image=registry.access.redhat.com/rhel7/pod-infrastructure:latest"
KUBELET_ARGS="--pod-infra-container-image=kubernetes/pause --cluster-domain=contoso.com --cluster-dns=172.16.3.1"

啟動(dòng)服務(wù)

$ for SERVICES in kube-proxy kubelet docker flanneld; do
    systemctl restart $SERVICES
    systemctl enable $SERVICES
    systemctl status $SERVICES 
done

在每個(gè)節(jié)點(diǎn)檢查

$ ip a | grep flannel | grep inet
會(huì)發(fā)現(xiàn)每個(gè)節(jié)點(diǎn)都有新的網(wǎng)絡(luò)接口docker0和flannel0

d181節(jié)點(diǎn)檢查應(yīng)可看到其他節(jié)點(diǎn)已經(jīng)啟動(dòng)
$ kubectl get nodes

第四部分：創(chuàng)建pods

以下操作在d181節(jié)點(diǎn)

編寫mysql配置文件

$ mkdir pods
$ cd pods
$ vim mysql.yaml
apiVersion: v1
kind: Pod
metadata:
  name: mysql-pod
  labels:
    name: mysql-pod
spec:
  containers:
    - resources:
        limits :
          cpu: 1
      image: mysql
      name: mysql
      env:
        - name: MYSQL_ROOT_PASSWORD
          value: 1a.genius
      ports:
        - containerPort: 3306
          name: mysql

創(chuàng)建pods

$ kubectl create -f mysql.yaml

測試

查看pod在哪個(gè)node上創(chuàng)建
$ kubectl get pods
$ kubectl get pod mysql -o wide
$ kubectl describe pod mysql

在上面查找到的節(jié)點(diǎn)上運(yùn)行
docker ps -a
注：發(fā)現(xiàn)有兩個(gè)2個(gè)容器泞辐，一個(gè)是mysql笔横，一個(gè)是pause，pause是Netowrk Container, 每啟動(dòng)一個(gè)Pod都會(huì)附加啟動(dòng)這樣一個(gè)容器咐吼，它的作用就只是簡單的等待吹缔，設(shè)置Pod的網(wǎng)絡(luò)

獲得pod更詳細(xì)的信息
kubectl get pod mysql -o yaml
kubectl get pod mysql -o json

手動(dòng)停止容器測試
docker stop $(docker ps -a -q)
注：發(fā)現(xiàn)k8s會(huì)自動(dòng)重新生成新容器，但如果關(guān)閉pod所在node再重啟該node锯茄，在d181上再查詢則發(fā)現(xiàn)pod會(huì)被銷毀厢塘，且不會(huì)自動(dòng)在其它node上創(chuàng)建新的pod

如何刪除pod
kubectl delete pod mysql-pod

第五部分：創(chuàng)建RC

注：創(chuàng)建rc時(shí)會(huì)自動(dòng)創(chuàng)建pod，所以運(yùn)行該命令前應(yīng)刪除上一步創(chuàng)建的pod
意義：Replication Controller可確保任何時(shí)候Kubernetes集群中有指定數(shù)量的pod副本

在d181節(jié)點(diǎn)運(yùn)行

編寫mysql rc文件

$ vim mysql-controller.yaml
apiVersion: v1
kind: ReplicationController
metadata:
   name: mysql-controller
   labels: 
     name: mysql-controller
spec:
  replicas: 2
  selector:
     name: mysql-pod
  template: 
    metadata:
     labels:
       name: mysql-pod
    spec:
      containers:
      - name: mysql
        image: mysql
        env:
        - name: MYSQL_ROOT_PASSWORD
          value: 1a.genius
        ports:
        - containerPort: 3306
          name: mysql

創(chuàng)建rc

kubectl create -f mysql-controller.yaml

查詢r(jià)c

kubectl get rc
kubectl get rc mysql-controller
kubectl describe rc mysql-controller

查詢pods

kubectl get pods
kubectl get pods -o wide
注：可見到在兩個(gè)node上創(chuàng)建了pod

擴(kuò)大副本數(shù)量

kubectl scale rc mysql-controller --replicas=3
kubectl get pods -o wide

減小副本數(shù)量

kubectl scale rc mysql-controller --replicas=1
kubectl get pods -o wide

刪除測試

刪除容器測試
docker rm <容器ID>
注：刪除某個(gè)容器肌幽，會(huì)馬上自動(dòng)啟動(dòng)新的容器

刪除rc不影響pod
kubectl delete rc mysql-controller

刪除rc及其下所有的pod
kubectl delete -f mysql-controller.yaml

第六部分：創(chuàng)建services（master節(jié)點(diǎn)）

注：創(chuàng)建service前最好已創(chuàng)建好pod
每個(gè)Pod都會(huì)分配一個(gè)單獨(dú)的IP地址晚碾，但這個(gè)IP地址會(huì)隨著Pod的銷毀而消失。Service解決通過一個(gè)IP訪問一組pod