手动搭建Kubernetes1.8高可用集群（5）Node

来源：互联网发布：空气净化器评测知乎编辑：程序博客网时间：2024/05/16 09:06

一、准备

1、etcd集群

2、Node2,Node3上搭建Node,以下所有操作都在Node3上进行。Node2只需要修改kubelet配置就可以了

3、创建目录，并分发证书

/etc/kubernetes/manifests    属主kube 属组kube-cert  权限0700/etc/kubernetes/ssl/etc/nginx

二、安装kubelet

1、复制二进制文件

docker run --rm -v /usr/local/bin:/systembindir quay.io/coreos/hyperkube:v1.8.3_coreos.0 /bin/cp /hyperkube /systembindir/kubelet

三、准备kubelet配置文件

1、/etc/systemd/system/kubelet.service

[Unit]Description=Kubernetes Kubelet ServerDocumentation=https://github.com/GoogleCloudPlatform/kubernetesAfter=docker.serviceWants=docker.socket[Service]EnvironmentFile=-/etc/kubernetes/kubelet.envExecStart=/usr/local/bin/kubelet \                $KUBE_LOGTOSTDERR \                $KUBE_LOG_LEVEL \                $KUBELET_API_SERVER \                $KUBELET_ADDRESS \                $KUBELET_PORT \                $KUBELET_HOSTNAME \                $KUBE_ALLOW_PRIV \                $KUBELET_ARGS \                $DOCKER_SOCKET \                $KUBELET_NETWORK_PLUGIN \                $KUBELET_CLOUDPROVIDERRestart=alwaysRestartSec=10s[Install]WantedBy=multi-user.target

2、/etc/kubernetes/kubelet.env

# logging to stderr means we get it in the systemd journalKUBE_LOGTOSTDERR="--logtostderr=true"KUBE_LOG_LEVEL="--v=2"# The address for the info server to serve on (set to 0.0.0.0 or "" for all interfaces)KUBELET_ADDRESS="--address=192.168.1.126 --node-ip=192.168.1.126"# The port for the info server to serve on# KUBELET_PORT="--port=10250"# You may leave this blank to use the actual hostnameKUBELET_HOSTNAME="--hostname-override=node3"KUBELET_ARGS="--pod-manifest-path=/etc/kubernetes/manifests \--cadvisor-port=0 \--pod-infra-container-image=gcr.io/google_containers/pause-amd64:3.0 \--node-status-update-frequency=10s \--docker-disable-shared-pid=True \--client-ca-file=/etc/kubernetes/ssl/ca.pem \--tls-cert-file=/etc/kubernetes/ssl/node-node3.pem \--tls-private-key-file=/etc/kubernetes/ssl/node-node3-key.pem \--anonymous-auth=false \--cgroup-driver=cgroupfs \--cgroups-per-qos=True \--fail-swap-on=False \--enforce-node-allocatable=""  --cluster-dns=10.233.0.3 --cluster-domain=cluster.local --resolv-conf=/etc/resolv.conf --kubeconfig=/etc/kubernetes/node-kubeconfig.yaml --require-kubeconfig --kube-reserved cpu=100m,memory=256M --node-labels=node-role.kubernetes.io/node=true  --feature-gates=Initializers=true,PersistentLocalVolumes=False  "KUBELET_NETWORK_PLUGIN="--network-plugin=cni --network-plugin-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin"# Should this cluster be allowed to run privileged docker containersKUBE_ALLOW_PRIV="--allow-privileged=true"KUBELET_CLOUDPROVIDER=""PATH=/usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin

3、/etc/kubernetes/node-kubeconfig.yaml

apiVersion: v1kind: Configclusters:- name: local  cluster:    certificate-authority: /etc/kubernetes/ssl/ca.pem    server: https://localhost:6443users:- name: kubelet  user:    client-certificate: /etc/kubernetes/ssl/node-node3.pem    client-key: /etc/kubernetes/ssl/node-node3-key.pemcontexts:- context:    cluster: local    user: kubelet  name: kubelet-cluster.localcurrent-context: kubelet-cluster.local

4、启动kubelet

systemctl start kubelet && systemctl enable kubelet

[root@node1 ~]# ss -tnlState       Recv-Q Send-Q                                                   Local Address:Port                                                                  Peer Address:Port              LISTEN      0      128                                                      192.168.1.123:10250                                                                            *:*                  LISTEN      0      128                                                      192.168.1.123:2379                                                                             *:*                  LISTEN      0      128                                                          127.0.0.1:2379                                                                             *:*                  LISTEN      0      128                                                      192.168.1.123:2380                                                                             *:*                  LISTEN      0      128                                                      192.168.1.123:10255                                                                            *:*                  LISTEN      0      128                                                                  *:22                                                                               *:*                  LISTEN      0      100                                                          127.0.0.1:25                                                                               *:*                  LISTEN      0      128                                                          127.0.0.1:10248                                                                            *:*                  LISTEN      0      128                                                                 :::22                                                                              :::*                  LISTEN      0      100                                                                ::1:25                                                                              :::*

四、配置kube-proxy,apiserver,scheduler,controller-manager

1、/etc/kubernetes/kube-proxy-kubeconfig.yaml

apiVersion: v1kind: Configclusters:- name: local  cluster:    certificate-authority: /etc/kubernetes/ssl/ca.pem    server: https://localhost:6443users:- name: kube-proxy  user:    client-certificate: /etc/kubernetes/ssl/kube-proxy-node3.pem    client-key: /etc/kubernetes/ssl/kube-proxy-node3-key.pemcontexts:- context:    cluster: local    user: kube-proxy  name: kube-proxy-cluster.localcurrent-context: kube-proxy-cluster.local

2、/etc/kubernetes/manifests/kube-proxy.manifest

apiVersion: v1kind: Podmetadata:  name: kube-proxy  namespace: kube-system  labels:    k8s-app: kube-proxy  annotations:    kubespray.kube-proxy-cert/serial: "DBA85609D00B0FAE"spec:  hostNetwork: true  dnsPolicy: ClusterFirst  containers:  - name: kube-proxy    image: quay.io/coreos/hyperkube:v1.8.3_coreos.0    imagePullPolicy: IfNotPresent    resources:      limits:        cpu: 500m        memory: 2000M      requests:        cpu: 150m        memory: 64M    command:    - /hyperkube    - proxy    - --v=2    - --kubeconfig=/etc/kubernetes/kube-proxy-kubeconfig.yaml    - --bind-address=192.168.1.126    - --cluster-cidr=10.233.64.0/18    - --proxy-mode=iptables    securityContext:      privileged: true    volumeMounts:    - mountPath: /etc/ssl/certs      name: ssl-certs-host      readOnly: true    - mountPath: "/etc/kubernetes/ssl"      name: etc-kube-ssl      readOnly: true    - mountPath: "/etc/kubernetes/kube-proxy-kubeconfig.yaml"      name: kubeconfig      readOnly: true    - mountPath: /var/run/dbus      name: var-run-dbus      readOnly: false  volumes:  - name: ssl-certs-host    hostPath:      path: /etc/pki/tls  - name: etc-kube-ssl    hostPath:      path: "/etc/kubernetes/ssl"  - name: kubeconfig    hostPath:      path: "/etc/kubernetes/kube-proxy-kubeconfig.yaml"  - name: var-run-dbus    hostPath:      path: /var/run/dbus

3、/etc/nginx/nginx.conf

error_log stderr notice;worker_processes auto;events {  multi_accept on;  use epoll;  worker_connections 1024;}stream {        upstream kube_apiserver {            least_conn;            server 192.168.1.121:6443;            server 192.168.1.122:6443;                    }        server {            listen        127.0.0.1:6443;            proxy_pass    kube_apiserver;            proxy_timeout 10m;            proxy_connect_timeout 1s;        }}

4、/etc/kubernetes/manifests/nginx-proxy.yml

apiVersion: v1kind: Podmetadata:  name: nginx-proxy  namespace: kube-system  labels:    k8s-app: kube-nginxspec:  hostNetwork: true  containers:  - name: nginx-proxy    image: nginx:1.11.4-alpine    imagePullPolicy: IfNotPresent    resources:      limits:        cpu: 300m        memory: 512M      requests:        cpu: 25m        memory: 32M    securityContext:      privileged: true    volumeMounts:    - mountPath: /etc/nginx      name: etc-nginx      readOnly: true  volumes:  - name: etc-nginx    hostPath:      path: /etc/nginx

四、验证

1、docker ps

[root@node3 ~]# docker psCONTAINER ID        IMAGE                                      COMMAND                  CREATED             STATUS              PORTS               NAMES6597dd5a1ce1        00bc1e841a8f                               "nginx -g 'daemon ..."   9 seconds ago       Up 9 seconds                            k8s_nginx-proxy_nginx-proxy-node3_kube-system_768ecc5f8a5c2500c7b1d97c4351756d_0fc03ac3c0887        gcr.io/google_containers/pause-amd64:3.0   "/pause"                 10 seconds ago      Up 9 seconds                            k8s_POD_nginx-proxy-node3_kube-system_768ecc5f8a5c2500c7b1d97c4351756d_06d7ae2b0e831        bd322856b660                               "/hyperkube proxy ..."   5 minutes ago       Up 5 minutes                            k8s_kube-proxy_kube-proxy-node3_kube-system_6e62ad1c50c542344a458bc75eef02f7_02a8382b3d714        gcr.io/google_containers/pause-amd64:3.0   "/pause"                 5 minutes ago       Up 5 minutes                            k8s_POD_kube-proxy-node3_kube-system_6e62ad1c50c542344a458bc75eef02f7_0c3befa316f36        quay.io/coreos/etcd:v3.2.4                 "/usr/local/bin/etcd"    About an hour ago   Up About an hour                        etcd3

2、ss -tnl

[root@node3 ~]# ss -tnlState       Recv-Q Send-Q                                                   Local Address:Port                                                                  Peer Address:Port              LISTEN      0      128                                                      192.168.1.125:10250                                                                            *:*                  LISTEN      0      128                                                          127.0.0.1:6443                                                                             *:*                  LISTEN      0      128                                                      192.168.1.125:2379                                                                             *:*                  LISTEN      0      128                                                          127.0.0.1:2379                                                                             *:*                  LISTEN      0      128                                                      192.168.1.125:2380                                                                             *:*                  LISTEN      0      128                                                      192.168.1.125:10255                                                                            *:*                  LISTEN      0      128                                                                  *:22                                                                               *:*                  LISTEN      0      100                                                          127.0.0.1:25                                                                               *:*                  LISTEN      0      128                                                          127.0.0.1:10248                                                                            *:*                  LISTEN      0      128                                                          127.0.0.1:10249                                                                            *:*                  LISTEN      0      128                                                                 :::10256                                                                           :::*                  LISTEN      0      128                                                                 :::22                                                                              :::*                  LISTEN      0      100                                                                ::1:25

3、kubectl get node #在Master上运行，除非你配置了

[root@node1 ~]# kubectl get nodeNAME      STATUS     ROLES         AGE       VERSIONnode1     NotReady   master        1h        v1.8.3+coreos.0node2     NotReady   master,node   1h        v1.8.3+coreos.0node3     NotReady   node          1m        v1.8.3+coreos.0

阅读全文

0 0