前言

  最近中国和印度的局势也是愈演愈烈。作为一个爱国青年我有些愤怒，但有时又及其的骄傲。不知道是因为中国外交强势还是软弱，怎样也应该有个态度吧？这是干嘛？就会抗议 在不就搞一些军演。有毛用啊？

  自己判断可能是国家有自己的打算吧！就好比狮子和疯狗一样何必那！中国和印度的纷纷扰扰，也不知道怎样霸气侧漏还是在伤仲永。

霸气侧漏是航母的电子弹射还是核潜艇或者是无人机.....

项目开始

我想大家都知道docker 但是也都玩过k8s吧！

搭建kubernetes集群时遇到一些问题，网上有不少搭建文档可以参考，但是满足以下网络互通才能算k8s集群ready。

需求如下：

k8s结构图如下：

以下是版本和机器信息:

节点初始化

• 更新CentOS-Base.repo为阿里云yum源

mv -f /etc/yum.repos.d/CentOS-Base.repo /etc/yum.repos.d/CentOS-Base.repo.bk; 
curl -o /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo

设置bridge

cat <<EOF > /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
net.bridge.bridge-nf-call-arptables = 1
EOF
sudo sysctl --system

• disable selinux （请不要用setenforce 0）

sed -i 's/SELINUX=enforcing/SELINUX=disabled/' /etc/selinux/config

• 关闭防火墙

sudo systemctl disable firewalld.service
sudo systemctl stop firewalld.service

• 关闭iptables

sudo yum install -y iptables-services;iptables -F;   #可略过sudo systemctl disable iptables.service
sudo systemctl stop iptables.service

• 安装相关软件

sudo yum install -y vim wget curl screen git etcd ebtables flannel
sudo yum install -y socat net-tools.x86_64 iperf bridge-utils.x86_64

• 安装docker (目前默认安装是1.12)

sudo yum install -y yum-utils device-mapper-persistent-data lvm2
sudo yum install -y libdevmapper* docker

• 安装kubernetes

• 方便复制粘贴如下：

  ##设置kubernetes.repo为阿里云源，适合国内cat <<EOF > /etc/yum.repos.d/kubernetes.repo [kubernetes] name=Kubernetes baseurl=https://mirrors.aliyun.com/kubernetes/yum/repos/kubernetes-el7-x86_64/ enabled=1 gpgcheck=0 EOF##设置kubernetes.repo为阿里云源，适合能连通google的网络cat <<EOF > /etc/yum.repos.d/kubernetes.repo [kubernetes] name=Kubernetes baseurl=http://yum.kubernetes.io/repos/kubernetes-el7-x86_64 enabled=1 gpgcheck=1 repo_gpgcheck=1 gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg    https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg EOF## 安装k8s 1.7.2 (kubernetes-cni会作为依赖一并安装,在此没有做版本指定)export K8SVERSION=1.7.2 sudo yum install -y "kubectl-${K8SVERSION}-0.x86_64" "kubelet-${K8SVERSION}-0.x86_64" "kubeadm-${K8SVERSION}-0.x86_64"
  

• 重启机器 （这一步是需要的）

reboot

---

重启机器后执行如下步骤

• 配置docker daemon并启动docker

cat <<EOF >/etc/sysconfig/docker
OPTIONS="-H unix:///var/run/docker.sock -H tcp://127.0.0.1:2375 --storage-driver=overlay --exec-opt native.cgroupdriver=cgroupfs --graph=/localdisk/docker/graph --insecure-registry=gcr.io --insecure-registry=quay.io  --insecure-registry=registry.cn-hangzhou.aliyuncs.com --registry-mirror=http://138f94c6.m.daocloud.io"EOF

systemctl start docker
systemctl status docker -l

• 拉取k8s 1.7.2 需要的镜像

quay.io/calico/node:v1.3.0
quay.io/calico/cni:v1.9.1
quay.io/calico/kube-policy-controller:v0.6.0

gcr.io/google_containers/pause-amd64:3.0
gcr.io/google_containers/kube-proxy-amd64:v1.7.2
gcr.io/google_containers/kube-apiserver-amd64:v1.7.2
gcr.io/google_containers/kube-controller-manager-amd64:v1.7.2
gcr.io/google_containers/kube-scheduler-amd64:v1.7.2
gcr.io/google_containers/etcd-amd64:3.0.17

gcr.io/google_containers/k8s-dns-dnsmasq-nanny-amd64:1.14.4
gcr.io/google_containers/k8s-dns-kube-dns-amd64:1.14.4
gcr.io/google_containers/k8s-dns-sidecar-amd64:1.14.4

• 在非k8s master节点 10.12.0.22 上启动ETCD (也可搭建成ETCD集群)

screen etcd -name="EtcdServer" -initial-advertise-peer-urls=http://10.12.0.22:2380 -listen-peer-urls=http://0.0.0.0:2380 -listen-client-urls=http://10.12.0.22:2379 -advertise-client-urls http://10.12.0.22:2379 -data-dir /var/lib/etcd/default.etcd

• 在每个节点上check是否可通达ETCD， 必须可通才行, 不通需要看下防火墙是不是没有关闭

etcdctl --endpoint=http://10.12.0.22:2379 member list
etcdctl --endpoint=http://10.12.0.22:2379 cluster-health

• 在k8s master节点上使用kubeadm启动，
  pod-ip网段设定为10.68.0.0/16, cluster-ip网段为默认10.96.0.0/16
  

  
  

• 如下命令在master节点上执行

cat << EOF >kubeadm_config.yaml
apiVersion: kubeadm.k8s.io/v1alpha1
kind: MasterConfiguration
api:
  advertiseAddress: 10.12.0.18  bindPort: 6443
etcd:
  endpoints:
  - http://10.12.0.22:2379
networking:
 dnsDomain: cluster.local
 serviceSubnet: 10.96.0.0/16
 podSubnet: 10.68.0.0/16
kubernetesVersion: v1.7.2#token: <string>#tokenTTL: 0EOF##kubeadm init --config kubeadm_config.yaml

• 执行kubeadm init命令后稍等几十秒，master上api-server, scheduler, controller-manager容器都启动起来，以下命令来check下master
  如下命令在master节点上执行

rm -rf $HOME/.kube
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

kubectl get cs -o wide --show-labels
kubectl get nodes -o wide --show-labels

• 节点加入, 需要kubeadm init命令输出的token, 如下命令在node节点上执行

systemctl start docker
systemctl start kubelet
kubeadm join --token *{6}.*{16} 10.12.0.18:6443 --skip-preflight-checks

• 在master节点上观察节点加入情况， 因为还没有创建网络，所以，所有master和node节点都是NotReady状态， kube-dns也是pending状态

kubectl get nodes -o wide
watch kubectl get all --all-namespaces -o wide

• 对calico.yaml做了修改
  删除ETCD创建部分，使用外部ETCD
  修改CALICO_IPV4POOL_CIDR为10.68.0.0/16
  calico.yaml如下：

• 
  

  
  

# Calico Version v2.3.0

# http://docs.projectcalico.org/v2.3/releases#v2.3.0

# This manifest includes the following component versions:

# calico/node:v1.3.0

# calico/cni:v1.9.1

# calico/kube-policy-controller:v0.6.0

# This Config Map is used to configure a self-hosted Calico installation.kind: Config MapapiVersion: v1metadata:  name: calico-config  namespace: kube-systemdata:  

# The location of your etcd cluster.  This uses the Service clusterIP defined below.  etcd_endpoints: "http://10.12.0.22:2379" 

# Configure the Calico backend to use.  calico_backend: "bird"  

# The CNI network configuration to install on each node.  cni_network_config: |-    {        "name": "k8s-pod-network",        "cniVersion": "0.1.0",        "type": "calico",        "etcd_endpoints": "__ETCD_ENDPOINTS__",        "log_level": "info",        "ipam": {        "type": "calico-ipam"        },        "policy": {      "type": "k8s",             "k8s_api_root": "https://__KUBERNETES_SERVICE_HOST__:__KUBERNETES_SERVICE_PORT__",             "k8s_auth_token": "__SERVICEACCOUNT_TOKEN__"        },  "kubernetes": {        "kubeconfig": "/etc/cni/net.d/__KUBECONFIG_FILENAME__"        }    }---

# This manifest installs the calico/node container, as well

# as the Calico CNI plugins and network config on

# each master and worker node in a Kubernetes cluster.kind: DaemonSetapiVersion: extensions/v1beta1metadata:  name: calico-node  namespace: kube-system  labels:   k8s-app: calico-nodespec:selector:matchLabels:k8s-app: calico-node template: metadata:labels: k8s-app: calico-node annotations:       

# Mark this pod as a critical add-on; when enabled, the critical add-on scheduler       

# reserves resources for critical add-on pods so that they can be rescheduled after        

# a failure.  This annotation works in tandem with the toleration below.        scheduler.alpha.kubernetes.io/critical-pod: ''  spec:      hostNetwork: true      tolerations: - key: node-role.kubernetes.io/master        effect: NoSchedule     

# Allow this pod to be rescheduled while the node is in "critical add-ons only" mode.      

# This, along with the annotation above marks this pod as a critical add-on.   key: CriticalAddonsOnly        operator: Exists  serviceAccountName: calico-cni-plugin     containers:        

# Runs calico/node container on each Kubernetes node.  This  # container programs network policy and routes on each       # host.- name: calico-node image:quay.io/calico/node:v1.3.0  env:            

# The location of the Calico etcd cluster.   - name: ETCD_ENDPOINTS              valueFrom:configMapKeyRef:        name: calico-config         key: etcd_endpoints            

# Enable BGP.  Disable to enforce policy only.            - name: CALICO_NETWORKING_BACKEND   valueFrom:config MapKeyRef: name: calico-config    key: calico_backend           

# Disable file logging so `kubectl logs` works.            - name: CALICO_DISABLE_FILE_LOGGING  value: "true"  

# Set Felix endpoint to host default action to ACCEPT.         - name: FELIX_DEFAULTENDPOINTTOHOSTACTION              value: "ACCEPT"           

# Configure the IP Pool from which Pod IPs will be chosen.            - name: CALICO_IPV4POOL_CIDR              value: "10.68.0.0/16"      - name: CALICO_IPV4POOL_IPIP              value: "always"          

# Disable IPv6 on Kubernetes.  - name: FELIX_IPV6SUPPORT      value: "false"            

# Set Felix logging to "info"  - name:FELIX_LOGSEVERITYSCREEN              value: "info"             

# Auto-detect the BGP IP address.  - name: IP  value: ""      securityContext:privileged: true  resources:  requests:      cpu: 250m     volumeMounts:        - mountPath: /lib/modules  name: lib-modules   readOnly: true    - mountP/var/run/calico             name: var-run-calico   readOnly: false        

# This container installs the Calico CNI binaries       

# and CNI network config file on each node.        - name: install-cni          image: quay.io/calico/cni:v1.9.1          command: ["/install-cni.sh"]          env:           

# The location of the Calico etcd cluster. - name: ETCD_ENDPOINTS              valueFrom:                configMapKeyRef:           name: calico-config               key: etcd_endpoints            

# The CNI network config to install on each node.            - name: CNI_NETWORK_CONFIG   valueFrom:                configMapKeyRef:    name: calico-config                  key: cni_network_config     volumeMounts:            - mountPath: /host/opt/cni/bin      ame: cni-bin-dir            - mountPath: /host/etc/cni/net.d              name: cni-net-dir      volumes:        

# Used by calico/node.        - name: lib-modules          hostPath:            path: /lib/modules        - name: var-run-calico          hostPath:            path: /var/run/calico        # Used to install CNI.        - name: cni-bin-dir          hostPath:            path: /opt/cni/bin        - name: cni-net-dir          hostPath:            path: /etc/cni/net.d---# This manifest deploys the Calico policy controller on Kubernetes.

# See https://github.com/projectcalico/k8s-policyapiVersion: extensions/v1beta1kind: Deploymentmetadata:  name: calico-policy-controller  namespace: kube-system  labels:    k8s-app: calico-policyspec:  

# The policy controller can only have a single active instance.  replicas: 1  strategy:    type: Recreate  template:    metadata:      name: calico-policy-controller      namespace: kube-system      labels:        k8s-app: calico-policy-controller      annotations:       

# Mark this pod as a critical add-on; when enabled, the critical add-on scheduler        

# reserves resources for critical add-on pods so that they can be rescheduled after        

# a failure.  This annotation works in tandem with the toleration below.        scheduler.alpha.kubernetes.io/critical-pod: ''    spec:     

# The policy controller must run in the host network namespace so that     

# it isn't governed by policy that would prevent it from working.      hostNetwork: true      tolerations:      - key: node-role.kubernetes.io/master        effect: NoSchedule      # Allow this pod to be rescheduled while the node is in "critical add-ons only" mode.      

# This, along with the annotation above marks this pod as a critical add-on.      - key: CriticalAddonsOnly        operator: Exists      serviceAccountName: calico-policy-controller      containers:        - name: calico-policy-controller          image: quay.io/calico/kube-policy-controller:v0.6.0          env:            

# The location of the Calico etcd cluster.            - name: ETCD_ENDPOINTS              valueFrom:                configMapKeyRef:                  name: calico-config                  key: etcd_endpoints           

# The location of the Kubernetes API.  Use the default Kubernetes            

# service for API access.            - name: K8S_API              value: "https://kubernetes.default:443"           

# Since we're running in the host namespace and might not have KubeDNS           

# access, configure the container's /etc/hosts to resolve           

# kubernetes.default to the correct service clusterIP.            - name: CONFIGURE_ETC_HOSTS              value: "true"---apiVersion: rbac.authorization.k8s.io/v1beta1kind: ClusterRoleBindingmetadata:  name: calico-cni-pluginroleRef:  apiGroup: rbac.authorization.k8s.io  kind: ClusterRole  name: calico-cni-pluginsubjects:- kind: ServiceAccount  name: calico-cni-plugin  namespace: kube-system---kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1beta1metadata:  name: calico-cni-plugin  namespace: kube-systemrules:  - apiGroups: [""]    resources:      - pods      - nodes    verbs:      - get---apiVersion: v1kind: ServiceAccountmetadata:  name: calico-cni-plugin  namespace: kube-system---apiVersion: rbac.authorization.k8s.io/v1beta1kind: ClusterRoleBindingmetadata:  name: calico-policy-controllerroleRef:  apiGroup: rbac.authorization.k8s.io  kind: ClusterRole  name: calico-policy-controllersubjects:- kind: ServiceAccount  name: calico-policy-controller  namespace: kube-system---kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1beta1metadata:  name: calico-policy-controller  namespace: kube-systemrules:  - apiGroups:    - ""    - extensions    resources:      - pods      - namespaces      - networkpolicies    verbs:      - watch      - list---apiVersion: v1kind: ServiceAccountmetadata:  name: calico-policy-controller  namespace: kube-system

• 创建calico跨主机网络, 在master节点上执行如下命令

kubectl apply -f calico.yaml

• 注意观察每个节点上会有名为calico-node-****的pod起来， calico-policy-controller和kube-dns也会起来， 这些pod都在kube-system名字空间里

>kubectl get all --all-namespaces

NAMESPACE     NAME                                                 READY     STATUS    RESTARTS   AGE
kube-system   po/calico-node-2gqf2                                 2/2       Running   0          19h
kube-system   po/calico-node-fg8gh                                 2/2       Running   0          19h
kube-system   po/calico-node-ksmrn                                 2/2       Running   0          19h
kube-system   po/calico-policy-controller-1727037546-zp4lp         1/1       Running   0          19h
kube-system   po/etcd-izuf6fb3vrfqnwbct6ivgwz                      1/1       Running   0          19h
kube-system   po/kube-apiserver-izuf6fb3vrfqnwbct6ivgwz            1/1       Running   0          19h
kube-system   po/kube-controller-manager-izuf6fb3vrfqnwbct6ivgwz   1/1       Running   0          19h
kube-system   po/kube-dns-2425271678-3t4g6                         3/3       Running   0          19h
kube-system   po/kube-proxy-6fg1l                                  1/1       Running   0          19h
kube-system   po/kube-proxy-fdbt2                                  1/1       Running   0          19h
kube-system   po/kube-proxy-lgf3z                                  1/1       Running   0          19h
kube-system   po/kube-scheduler-izuf6fb3vrfqnwbct6ivgwz            1/1       Running   0          19h

NAMESPACE     NAME                       CLUSTER-IP      EXTERNAL-IP   PORT(S)         AGE
default       svc/kubernetes             10.96.0.1       <none>        443/TCP         19h
kube-system   svc/kube-dns               10.96.0.10      <none>        53/UDP,53/TCP   19h


NAMESPACE     NAME                              DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
kube-system   deploy/calico-policy-controller   1         1         1            1           19h
kube-system   deploy/kube-dns                   1         1         1            1           19h


NAMESPACE     NAME                                     DESIRED   CURRENT   READY     AGE
kube-system   rs/calico-policy-controller-1727037546   1         1         1         19h
kube-system   rs/kube-dns-2425271678                   1         1         1         19h

• 部署dash-board

wget https://raw.githubusercontent.com/kubernetes/dashboard/master/src/deploy/kubernetes-dashboard.yaml
kubectl create -f kubernetes-dashboard.yaml

• 部署heapster

wget https://github.com/kubernetes/heapster/archive/v1.4.0.tar.gz
tar -zxvf v1.4.0.tar.gzcd heapster-1.4.0/deploy/kube-config/influxdb
kubectl create -f ./

其他命令

• 强制删除某个pod

kubectl delete pod <podname> --namespace=<namspacer>  --grace-period=0 --force

• 重置某个node节点

kubeadm reset 
systemctl stop kubelet;
docker ps -aq | xargs docker rm -fv
find /var/lib/kubelet | xargs -n 1 findmnt -n -t tmpfs -o TARGET -T | uniq | xargs -r umount -v;
rm -rf /var/lib/kubelet /etc/kubernetes/ /var/lib/etcd 
systemctl start kubelet;

• 访问dashboard (在master节点上执行)

kubectl proxy --address=0.0.0.0 --port=8001 --accept-hosts='^.*'
or
kubectl proxy --port=8011 --address=192.168.61.100 --accept-hosts='^192\.168\.61\.*'

access to http://0.0.0.0:8001/ui

• Access to API with authentication token

APISERVER=$(kubectl config view | grep server | cut -f 2- -d ":" | tr -d " ")
TOKEN=$(kubectl describe secret $(kubectl get secrets | grep default | cut -f1 -d ' ') | grep -E '^token' | cut -f2 -d':' | tr -d '\t')
curl $APISERVER/api --header "Authorization: Bearer $TOKEN" --insecure

• 让master节点参与调度，默认master是不参与到任务调度中的

kubectl taint nodes --all node-role.kubernetes.io/master-
or
kubectl taint nodes --all dedicated-

• kubernetes master 消除隔离之前 Annotations

Name:           izuf6fb3vrfqnwbct6ivgwzRole:Labels:         beta.kubernetes.io/arch=amd64
            beta.kubernetes.io/os=linux
            kubernetes.io/hostname=izuf6fb3vrfqnwbct6ivgwz
            node-role.kubernetes.io/master=Annotations:        node.alpha.kubernetes.io/ttl=0
            volumes.kubernetes.io/controller-managed-attach-detach=true

• kubernetes master 消除隔离之后 Annotations

Name:           izuf6fb3vrfqnwbct6ivgwzRole:Labels:         beta.kubernetes.io/arch=amd64
            beta.kubernetes.io/os=linux
            kubernetes.io/hostname=izuf6fb3vrfqnwbct6ivgwz
            node-role.kubernetes.io/master=Annotations:        node.alpha.kubernetes.io/ttl=0
            volumes.kubernetes.io/controller-managed-attach-detach=trueTaints:         <none>

总结：通过测试已经完成但是还有错看过文档的伙伴能猜到吗？