近来容器对企业来说已经不是什么陌生的概念,Kubernetes作为Google开源的容器运行平台,受到了大家的热捧。搭建一套完整的kubernetes平台,也成为试用这套平台必须迈过的坎儿。kubernetes1.5版本以及之前,安装还是相对比较方便的,官方就有通过yum源在centos7安装kubernetes。但是在kubernetes1.6之后,安装就比较繁琐了,需要证书各种认证,对于刚接触kubernetes的人来说很不友好,按照官方文档在本地安装“集群”的的话,我觉得你肯定是跑不起来的,除非你突破了GFW的限制,还要懂得怎么样不断修改参数。
令我惊讶的事,漫天都在吹捧kubernetes的强大,居然一篇可以帮助安装的文档都找不到(你可以说我菜,但是我找的官方和非官方文档不下十几篇,无一成功)。
我自己搞了一个星期,把kubernetes1.7.3版本(网络组件选用Calico),终于按照官方文档+填坑的方式部署成功,写此文希望能帮助更多的人部署kubernetes1.7成功。
[root@centos7-base-ok]# cat /etc/redhat-release
CentOS Linux release 7.3.1611 (Core)
[root@centos7-base-ok]# cat /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
k8s-1 192.168.80.28
k8s-2 192.168.80.35
k8s-3 192.168.80.14
注意:现在docker已经更新到CE版本,但是kubernetes官方文档说在1.12上测试通过,最近版本的兼容性未测试,为了避免后面出现大坑,我们还是乖乖安装1.12版本的docker。
1.新建docker.repo文件,将文件移动到/etc/yum.repos.d/目录下
[root@centos7-base-ok]# cat /etc/yum.repos.d/docker.repo
[dockerrepo]
name=Docker Repository
baseurl=https://yum.dockerproject.org/repo/main/centos/7/
enabled=1
gpgcheck=1
gpgkey=https://yum.dockerproject.org/gpg
2.运行yum命令,找到需要安装的docker版本
10:21 [root@centos7-base-ok]# yum list|grep docker | sort -r
python2-avocado-plugins-runner-docker.noarch
python-dockerpty.noarch 0.4.1-6.el7 epel
python-dockerfile-parse.noarch 0.0.5-1.el7 epel
python-docker-scripts.noarch 0.4.4-1.el7 epel
python-docker-pycreds.noarch 1.10.6-1.el7 extras
python-docker-py.noarch 1.10.6-1.el7 extras
kdocker.x86_64 4.9-1.el7 epel
golang-github-fsouza-go-dockerclient-devel.x86_64
docker.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-v1.10-migrator.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-unit-test.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-registry.x86_64 0.9.1-7.el7 extras
docker-registry.noarch 0.6.8-8.el7 extras
docker-python.x86_64 1.4.0-115.el7 extras
docker-novolume-plugin.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-lvm-plugin.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-logrotate.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-latest.x86_64 1.13.1-13.gitb303bf6.el7.centos
docker-latest-v1.10-migrator.x86_64 1.13.1-13.gitb303bf6.el7.centos
docker-latest-logrotate.x86_64 1.13.1-13.gitb303bf6.el7.centos
docker-forward-journald.x86_64 1.10.3-44.el7.centos extras
docker-engine.x86_64 17.05.0.ce-1.el7.centos dockerrepo
docker-engine.x86_64 1.12.6-1.el7.centos @dockerrepo
docker-engine-selinux.noarch 17.05.0.ce-1.el7.centos @dockerrepo
docker-engine-debuginfo.x86_64 17.05.0.ce-1.el7.centos dockerrepo
docker-distribution.x86_64 2.6.1-1.el7 extras
docker-devel.x86_64 1.3.2-4.el7.centos extras
docker-compose.noarch 1.9.0-5.el7 epel
docker-common.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-client.x86_64 2:1.12.6-32.git88a4867.el7.centos
docker-client-latest.x86_64 1.13.1-13.gitb303bf6.el7.centos
cockpit-docker.x86_64 141-3.el7.centos extras
3.找到对应版本后,执行yum install -y 包名+版本号,安装1.12版本的docker-engine
[root@centos7-base-ok]# yum install -y docker-engine.x86_64-1.12.6-1.el7.centos
4.执行docker version命令,验证docker安装版本,执行docker run命令,验证docker是否安装成功
[root@centos7-base-ok]# docker version
Client:
Version: 1.12.6
API version: 1.24
Go version: go1.6.4
Git commit: 78d1802
Built: Tue Jan 10 20:20:01 2017
OS/Arch: linux/amd64
Server:
Version: 1.12.6
API version: 1.24
Go version: go1.6.4
Git commit: 78d1802
Built: Tue Jan 10 20:20:01 2017
OS/Arch: linux/amd64
5.设置开机启动,启动容器,docker安装完成
[root@centos7-base-ok]# systemctl enbale docker && systemctl start docker
注意:此步骤是填坑的开始,因为官方文档的yum源在国内无法使用,安装完成后注意观察你的/var/log/message日志,会疯狂报错,别着急,跟着我一步一步来填坑。
1.新建kubernetes.repo文件,将文件移动到/etc/yum.repos.d/目录下(所有节点)
[root@centos7-base-ok]# cat /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
2.通过yum安装kubectl、kubelet、kubeadm(所有节点)
[root@centos7-base-ok]# cat /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
3.修改kubelet配置,启动kubelet(所有节点)
注意:时刻查看/var/log/message的日志输出,会看到kubelet一直启动失败。
编辑10-kubeadm.conf的文件,修改cgroup-driver配置:
[root@centos7-base-ok]# cat /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
[Service]
Environment="KUBELET_KUBECONFIG_ARGS=--kubeconfig=/etc/kubernetes/kubelet.conf --require-kubeconfig=true"
Environment="KUBELET_SYSTEM_PODS_ARGS=--pod-manifest-path=/etc/kubernetes/manifests --allow-privileged=true"
Environment="KUBELET_NETWORK_ARGS=--network-plugin=cni --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin"
Environment="KUBELET_DNS_ARGS=--cluster-dns=10.96.0.10 --cluster-domain=cluster.local"
Environment="KUBELET_AUTHZ_ARGS=--authorization-mode=Webhook --client-ca-file=/etc/kubernetes/pki/ca.crt"
Environment="KUBELET_CADVISOR_ARGS=--cadvisor-port=0"
Environment="KUBELET_CGROUP_ARGS=--cgroup-driver=cgroupfs"
ExecStart=
ExecStart=/usr/bin/kubelet $KUBELET_KUBECONFIG_ARGS $KUBELET_SYSTEM_PODS_ARGS $KUBELET_NETWORK_ARGS $KUBELET_DNS_ARGS $KUBELET_AUTHZ_ARGS $KUBELET_CADVISOR_ARGS $KUBELET_CGROUP_ARGS $KUBELET_EXTRA_ARGS
将“--cgroup-driver=systems”修改成为“--cgroup-driver=cgroupfs”,重新启动kubelet。
[root@centos7-base-ok]# systemctl restart kubelet
4.下载安装k8s依赖镜像
注意:此步骤非常关键,kubenetes初始化启动会依赖这些镜像,天朝的网络肯定是拉不下来google的镜像的,一般人过了上一关,这一关未必过的去,一定要提前把镜像下载到本地,kubeadm安装才会继续,下面我会列出来master节点和node依赖的镜像列表。(备注:考虑到随着kubernetes版本一直更新,镜像也可能会有变化,大家可以先执行
kubeadm init
生成配置文件,日志输出到[apiclient] Created API client, waiting for the control plane to become ready
这一行就会卡住不动了,你可以直接执行ctrl + c
中止命令执行,然后查看ls -ltr /etc/kubernetes/manifests/
yaml文件列表,每个文件都会写着镜像的地址和版本)
在这里我提一个可以解决下载google镜像的方法,就是买一台可以下载的机器,安装代理软件,在需要下载google镜像的机器的docker设置HTTP_PROXY
配置项,配置好自己的服务代理即可(也可以直接买可以访问到google的服务器安装).
master节点:
REPOSITORY TAG IMAGE ID CREATED SIZE
quay.io/calico/kube-policy-controller v0.7.0 fe3174230993 3 days ago 21.94 MB
kubernetesdashboarddev/kubernetes-dashboard-amd64 head e2cadb73b2df 5 days ago 136.5 MB
quay.io/calico/node v2.4.1 7643422fdf0f 6 days ago 277.4 MB
gcr.io/google_containers/kube-controller-manager-amd64 v1.7.3 d014f402b272 11 days ago 138 MB
gcr.io/google_containers/kube-apiserver-amd64 v1.7.3 a1cc3a3d8d0d 11 days ago 186.1 MB
gcr.io/google_containers/kube-scheduler-amd64 v1.7.3 51967bf607d3 11 days ago 77.2 MB
gcr.io/google_containers/kube-proxy-amd64 v1.7.3 54d2a8698e3c 11 days ago 114.7 MB
quay.io/calico/cni v1.10.0 88ca805c8ddd 13 days ago 70.25 MB
gcr.io/google_containers/kubernetes-dashboard-amd64 v1.6.3 691a82db1ecd 2 weeks ago 139 MB
quay.io/coreos/etcd v3.1.10 47bb9dd99916 4 weeks ago 34.56 MB
gcr.io/google_containers/k8s-dns-sidecar-amd64 1.14.4 38bac66034a6 7 weeks ago 41.81 MB
gcr.io/google_containers/k8s-dns-kube-dns-amd64 1.14.4 a8e00546bcf3 7 weeks ago 49.38 MB
gcr.io/google_containers/k8s-dns-dnsmasq-nanny-amd64 1.14.4 f7f45b9cb733 7 weeks ago 41.41 MB
gcr.io/google_containers/etcd-amd64 3.0.17 243830dae7dd 5 months ago 168.9 MB
gcr.io/google_containers/pause-amd64 3.0 99e59f495ffa 15 months ago 746.9 kB
node节点:
[root@centos7-base-ok]# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
kubernetesdashboarddev/kubernetes-dashboard-amd64 head e2cadb73b2df 5 days ago 137MB
quay.io/calico/node v2.4.1 7643422fdf0f 6 days ago 277MB
gcr.io/google_containers/kube-proxy-amd64 v1.7.3 54d2a8698e3c 11 days ago 115MB
quay.io/calico/cni v1.10.0 88ca805c8ddd 13 days ago 70.3MB
gcr.io/google_containers/kubernetes-dashboard-amd64 v1.6.3 691a82db1ecd 2 weeks ago 139MB
nginx latest b8efb18f159b 2 weeks ago 107MB
hello-world latest 1815c82652c0 2 months ago 1.84kB
gcr.io/google_containers/pause-amd64 3.0 99e59f495ffa 15 months ago 747kB
5.利用kubeadm初始化服务(master节点)
注意:如果你在上一步执行过
kubeadm init
命令,没有关系,此步执行只需要执行时加上--skip-preflight-checks
这个配置项即可。
注意:执行
kubeadm init
的--pod-network-cidr
参数和选择的网络组件有关系,详细可以看官方文档说明,本文选用的网络组件为 Calico
[root@centos7-base-ok]# kubeadm init --pod-network-cidr=192.168.0.0/16 --apiserver-advertise-address=0.0.0.0 --apiserver-cert-extra-sans=192.168.80.28,192.168.80.14,192.168.80.35,127.0.0.1,k8s-1,k8s-2,k8s-3,192.168.0.1 --skip-preflight-checks
参数说明:
参数名称 | 必选 | 参数说明 |
---|---|---|
pod-network-cidr | Yes | For certain networking solutions the Kubernetes master can also play a role in allocating network ranges (CIDRs) to each node. This includes many cloud providers and flannel. You can specify a subnet range that will be broken down and handed out to each node with the --pod-network-cidr flag. This should be a minimum of a /16 so controller-manager is able to assign /24 subnets to each node in the cluster. If you are using flannel with this manifest you should use --pod-network-cidr=10.244.0.0/16. Most CNI based networking solutions do not require this flag. |
apiserver-advertise-address | Yes | This is the address the API Server will advertise to other members of the cluster. This is also the address used to construct the suggested kubeadm join line at the end of the init process. If not set (or set to 0.0.0.0) then IP for the default interface will be used. |
apiserver-cert-extra-sans | Yes | Additional hostnames or IP addresses that should be added to the Subject Alternate Name section for the certificate that the API Server will use. If you expose the API Server through a load balancer and public DNS you could specify this with. |
其它的 kubeadm
参数设置请参照 官方文档
6.做一枚安静的美男子,等待安装成功,安装成功后你会看到日志如下(master节点):
注意:记录这段日志,后面添加node节点要用到。
[apiclient] All control plane components are healthy after 22.003243 seconds
[token] Using token: 33729e.977f7b5d0a9b5f3e
[apiconfig] Created RBAC rules
[addons] Applied essential addon: kube-proxy
[addons] Applied essential addon: kube-dns
Your Kubernetes master has initialized successfully!
To start using your cluster, you need to run (as a regular user):
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
http://kubernetes.io/docs/admin/addons/
You can now join any number of machines by running the following on each node
as root:
kubeadm join --token xxxxxxx 192.168.80.28:6443
7.创建kube的目录,添加kubectl配置(master节点)
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
8.用 kubectl
添加网络组件Calico(master节点)
kubectl apply -f http://docs.projectcalico.org/v2.4/getting-started/kubernetes/installation/hosted/kubeadm/1.6/calico.yaml
注意:此处坑为该文件未必下载的到,建议还是提前下载到本地,然后执行
kubectl apply -f <本地路径>
9.确认安装是否成功(master节点)
9.1 打开你的/var/log/messages,查看是否有报错,理论上,执行完上一步过去5分钟,日志应该不会有任何错误出现,如果持续报错,并且过了10分钟错误依然没有消失,检查之前的步骤是否有问题
9.2 运行 kubectl get pods --all-namespaces
查看结果,如果STATUS都为Running,恭喜你,你的master已经安装成功了。
注意:你的结果显示的条数未必和我完全一样,因为我这里有node节点的相关信息,而你还没有添加node节点。
[root@centos7-base-ok]# kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
default nginx-app-1666850838-4z2tb 1/1 Running 0 3d
kube-system calico-etcd-0ssdd 1/1 Running 0 3d
kube-system calico-node-1zfxd 2/2 Running 1 3d
kube-system calico-node-s2gfs 2/2 Running 1 3d
kube-system calico-node-xx30v 2/2 Running 1 3d
kube-system calico-policy-controller-336633499-wgl8j 1/1 Running 0 3d
kube-system etcd-k8s-1 1/1 Running 0 3d
kube-system kube-apiserver-k8s-1 1/1 Running 0 3d
kube-system kube-controller-manager-k8s-1 1/1 Running 0 3d
kube-system kube-dns-2425271678-trmxx 3/3 Running 1 3d
kube-system kube-proxy-79kkh 1/1 Running 0 3d
kube-system kube-proxy-n1g6j 1/1 Running 0 3d
kube-system kube-proxy-vccr6 1/1 Running 0 3d
kube-system kube-scheduler-k8s-1 1/1 Running 0 3d
10.安装node节点,执行在master节点执行成功输出的日志语句(node节点执行)
注意:执行如下语句的之前,一定要确认node节点下载了上文提到的镜像,否则因为镜像下载不成功会导致node节点初始化失败;第二点,一定要时刻查看/var/log/messages日志,如果镜像版本发生变化,在日志里会提示需要下载的镜像;第三点,就是要有耐心,如果你的网络可以下载到镜像,你当个安静的美男子就可以了,因为
kubeadm
会帮你做一切,知道你发现/var/log/messages不再有错误日志出现,说明它已经帮你搞定了所有事情,你可以开心的玩耍了。
[root@centos7-base-ok]# kubeadm join --token xxxxxxxx 192.168.80.28:6443
kubectl get nodes
查看节点状态。注意:node的状态会变化,添加成功后才是Ready。
[root@centos7-base-ok]# kubectl get nodes
NAME STATUS AGE VERSION
k8s-1 Ready 3d v1.7.3
k8s-2 Ready 3d v1.7.3
k8s-3 Ready 3d v1.7.3
12.恭喜你,你可以开心的进行kubernetes1.7.3之旅了
Kubernetes,想说爱你不容易啊 ,欢迎其它团队或者个人与我们团队进行交流,有意向可以评论区给我留言。
补充:目前官方说dashboard的HEAD版本支持1.7,但是我试了下dashboard确实不行,希望官方加快修复,还有就是多些错误定位的方法,否则很难提出具体的问题。