• 部署master节点
    • TLS 证书文件
    • 下载最新版本的二进制文件
    • 配置和启动 kube-apiserver
    • 配置和启动 kube-controller-manager
      • 启动 kube-controller-manager
    • 配置和启动 kube-scheduler
      • 启动 kube-scheduler
    • 验证 master 节点功能

    部署master节点

    kubernetes master 节点包含的组件:

    • kube-apiserver
    • kube-scheduler
    • kube-controller-manager

    目前这三个组件需要部署在同一台机器上。

    • kube-schedulerkube-controller-managerkube-apiserver 三者的功能紧密相关;
    • 同时只能有一个 kube-schedulerkube-controller-manager 进程处于工作状态,如果运行多个,则需要通过选举产生一个 leader;

    • 暂时未实现master节点的高可用
    • master节点上没有部署flannel网络插件,如果想要在master节点上也能访问ClusterIP,请参考下一节部署node节点中的配置Flanneld部分。

    TLS 证书文件

    以下pem证书文件我们在创建TLS证书和秘钥这一步中已经创建过了,token.csv文件在创建kubeconfig文件的时候创建。我们再检查一下。

    1. $ ls /etc/kubernetes/ssl
    2. admin-key.pem admin.pem ca-key.pem ca.pem kube-proxy-key.pem kube-proxy.pem kubernetes-key.pem kubernetes.pem

    下载最新版本的二进制文件

    有两种下载方式,请注意下载对应的Kubernetes版本。

    方式一

    从 github release 页面 下载发布版 tarball,解压后再执行下载脚本

    1. wget https://github.com/kubernetes/kubernetes/releases/download/v1.6.0/kubernetes.tar.gz
    2. tar -xzvf kubernetes.tar.gz
    3. cd kubernetes
    4. ./cluster/get-kube-binaries.sh

    方式二

    CHANGELOG页面 下载 clientserver tarball 文件

    server 的 tarball kubernetes-server-linux-amd64.tar.gz 已经包含了 client(kubectl) 二进制文件,所以不用单独下载kubernetes-client-linux-amd64.tar.gz文件;

    1. # wget https://dl.k8s.io/v1.6.0/kubernetes-client-linux-amd64.tar.gz
    2. wget https://dl.k8s.io/v1.6.0/kubernetes-server-linux-amd64.tar.gz
    3. tar -xzvf kubernetes-server-linux-amd64.tar.gz
    4. cd kubernetes
    5. tar -xzvf kubernetes-src.tar.gz

    将二进制文件拷贝到指定路径

    1. cp -r server/bin/{kube-apiserver,kube-controller-manager,kube-scheduler,kubectl,kube-proxy,kubelet} /usr/local/bin/

    配置和启动 kube-apiserver

    创建 kube-apiserver的service配置文件

    service配置文件/usr/lib/systemd/system/kube-apiserver.service内容:

    1. [Unit]
    2. Description=Kubernetes API Service
    3. Documentation=https://github.com/GoogleCloudPlatform/kubernetes
    4. After=network.target
    5. After=etcd.service
    6. [Service]
    7. EnvironmentFile=-/etc/kubernetes/config
    8. EnvironmentFile=-/etc/kubernetes/apiserver
    9. ExecStart=/usr/local/bin/kube-apiserver \
    10. $KUBE_LOGTOSTDERR \
    11. $KUBE_LOG_LEVEL \
    12. $KUBE_ETCD_SERVERS \
    13. $KUBE_API_ADDRESS \
    14. $KUBE_API_PORT \
    15. $KUBELET_PORT \
    16. $KUBE_ALLOW_PRIV \
    17. $KUBE_SERVICE_ADDRESSES \
    18. $KUBE_ADMISSION_CONTROL \
    19. $KUBE_API_ARGS
    20. Restart=on-failure
    21. Type=notify
    22. LimitNOFILE=65536
    23. [Install]
    24. WantedBy=multi-user.target

    /etc/kubernetes/config文件的内容为:

    1. ###
    2. # kubernetes system config
    3. #
    4. # The following values are used to configure various aspects of all
    5. # kubernetes services, including
    6. #
    7. # kube-apiserver.service
    8. # kube-controller-manager.service
    9. # kube-scheduler.service
    10. # kubelet.service
    11. # kube-proxy.service
    12. # logging to stderr means we get it in the systemd journal
    13. KUBE_LOGTOSTDERR="--logtostderr=true"
    14. # journal message level, 0 is debug
    15. KUBE_LOG_LEVEL="--v=0"
    16. # Should this cluster be allowed to run privileged docker containers
    17. KUBE_ALLOW_PRIV="--allow-privileged=true"
    18. # How the controller-manager, scheduler, and proxy find the apiserver
    19. #KUBE_MASTER="--master=http://test-001.jimmysong.io:8080"
    20. KUBE_MASTER="--master=http://172.20.0.113:8080"

    该配置文件同时被kube-apiserver、kube-controller-manager、kube-scheduler、kubelet、kube-proxy使用。

    apiserver配置文件/etc/kubernetes/apiserver内容为:

    1. ###
    2. ## kubernetes system config
    3. ##
    4. ## The following values are used to configure the kube-apiserver
    5. ##
    6. #
    7. ## The address on the local server to listen to.
    8. #KUBE_API_ADDRESS="--insecure-bind-address=test-001.jimmysong.io"
    9. KUBE_API_ADDRESS="--advertise-address=172.20.0.113 --bind-address=172.20.0.113 --insecure-bind-address=172.20.0.113"
    10. #
    11. ## The port on the local server to listen on.
    12. #KUBE_API_PORT="--port=8080"
    13. #
    14. ## Port minions listen on
    15. #KUBELET_PORT="--kubelet-port=10250"
    16. #
    17. ## Comma separated list of nodes in the etcd cluster
    18. KUBE_ETCD_SERVERS="--etcd-servers=https://172.20.0.113:2379,https://172.20.0.114:2379,https://172.20.0.115:2379"
    19. #
    20. ## Address range to use for services
    21. KUBE_SERVICE_ADDRESSES="--service-cluster-ip-range=10.254.0.0/16"
    22. #
    23. ## default admission control policies
    24. KUBE_ADMISSION_CONTROL="--admission-control=ServiceAccount,NamespaceLifecycle,NamespaceExists,LimitRanger,ResourceQuota"
    25. #
    26. ## Add your own!
    27. KUBE_API_ARGS="--authorization-mode=RBAC --runtime-config=rbac.authorization.k8s.io/v1beta1 --kubelet-https=true --experimental-bootstrap-token-auth --token-auth-file=/etc/kubernetes/token.csv --service-node-port-range=30000-32767 --tls-cert-file=/etc/kubernetes/ssl/kubernetes.pem --tls-private-key-file=/etc/kubernetes/ssl/kubernetes-key.pem --client-ca-file=/etc/kubernetes/ssl/ca.pem --service-account-key-file=/etc/kubernetes/ssl/ca-key.pem --etcd-cafile=/etc/kubernetes/ssl/ca.pem --etcd-certfile=/etc/kubernetes/ssl/kubernetes.pem --etcd-keyfile=/etc/kubernetes/ssl/kubernetes-key.pem --enable-swagger-ui=true --apiserver-count=3 --audit-log-maxage=30 --audit-log-maxbackup=3 --audit-log-maxsize=100 --audit-log-path=/var/lib/audit.log --event-ttl=1h"
    • --experimental-bootstrap-token-auth Bootstrap Token Authentication在1.9版本已经变成了正式feature,参数名称改为--enable-bootstrap-token-auth
    • 如果中途修改过--service-cluster-ip-range地址,则必须将default命名空间的kubernetes的service给删除,使用命令:kubectl delete service kubernetes,然后系统会自动用新的ip重建这个service,不然apiserver的log有报错the cluster IP x.x.x.x for service kubernetes/default is not within the service CIDR x.x.x.x/16; please recreate
    • --authorization-mode=RBAC 指定在安全端口使用 RBAC 授权模式,拒绝未通过授权的请求;
    • kube-scheduler、kube-controller-manager 一般和 kube-apiserver 部署在同一台机器上,它们使用非安全端口和 kube-apiserver通信;
    • kubelet、kube-proxy、kubectl 部署在其它 Node 节点上,如果通过安全端口访问 kube-apiserver,则必须先通过 TLS 证书认证,再通过 RBAC 授权;
    • kube-proxy、kubectl 通过在使用的证书里指定相关的 User、Group 来达到通过 RBAC 授权的目的;
    • 如果使用了 kubelet TLS Boostrap 机制,则不能再指定 --kubelet-certificate-authority--kubelet-client-certificate--kubelet-client-key 选项,否则后续 kube-apiserver 校验 kubelet 证书时出现 ”x509: certificate signed by unknown authority“ 错误;
    • --admission-control 值必须包含 ServiceAccount
    • --bind-address 不能为 127.0.0.1
    • runtime-config配置为rbac.authorization.k8s.io/v1beta1,表示运行时的apiVersion;
    • --service-cluster-ip-range 指定 Service Cluster IP 地址段,该地址段不能路由可达;
    • 缺省情况下 kubernetes 对象保存在 etcd /registry 路径下,可以通过 --etcd-prefix 参数进行调整;
    • 如果需要开通http的无认证的接口,则可以增加以下两个参数:--insecure-port=8080 --insecure-bind-address=127.0.0.1。注意,生产上不要绑定到非127.0.0.1的地址上

    Kubernetes 1.9

    • 对于Kubernetes1.9集群,需要注意配置KUBE_API_ARGS环境变量中的--authorization-mode=Node,RBAC,增加对Node授权的模式,否则将无法注册node。
    • --experimental-bootstrap-token-auth Bootstrap Token Authentication在kubernetes 1.9版本已经废弃,参数名称改为--enable-bootstrap-token-auth

    完整 unit 见 kube-apiserver.service

    启动kube-apiserver

    1. systemctl daemon-reload
    2. systemctl enable kube-apiserver
    3. systemctl start kube-apiserver
    4. systemctl status kube-apiserver

    配置和启动 kube-controller-manager

    创建 kube-controller-manager的serivce配置文件

    文件路径/usr/lib/systemd/system/kube-controller-manager.service

    1. [Unit]
    2. Description=Kubernetes Controller Manager
    3. Documentation=https://github.com/GoogleCloudPlatform/kubernetes
    4. [Service]
    5. EnvironmentFile=-/etc/kubernetes/config
    6. EnvironmentFile=-/etc/kubernetes/controller-manager
    7. ExecStart=/usr/local/bin/kube-controller-manager \
    8. $KUBE_LOGTOSTDERR \
    9. $KUBE_LOG_LEVEL \
    10. $KUBE_MASTER \
    11. $KUBE_CONTROLLER_MANAGER_ARGS
    12. Restart=on-failure
    13. LimitNOFILE=65536
    14. [Install]
    15. WantedBy=multi-user.target

    配置文件/etc/kubernetes/controller-manager

    1. ###
    2. # The following values are used to configure the kubernetes controller-manager
    3. # defaults from config and apiserver should be adequate
    4. # Add your own!
    5. KUBE_CONTROLLER_MANAGER_ARGS="--address=127.0.0.1 --service-cluster-ip-range=10.254.0.0/16 --cluster-name=kubernetes --cluster-signing-cert-file=/etc/kubernetes/ssl/ca.pem --cluster-signing-key-file=/etc/kubernetes/ssl/ca-key.pem --service-account-private-key-file=/etc/kubernetes/ssl/ca-key.pem --root-ca-file=/etc/kubernetes/ssl/ca.pem --leader-elect=true"
    • --service-cluster-ip-range 参数指定 Cluster 中 Service 的CIDR范围,该网络在各 Node 间必须路由不可达,必须和 kube-apiserver 中的参数一致;
    • --cluster-signing-* 指定的证书和私钥文件用来签名为 TLS BootStrap 创建的证书和私钥;
    • --root-ca-file 用来对 kube-apiserver 证书进行校验,指定该参数后,才会在Pod 容器的 ServiceAccount 中放置该 CA 证书文件
    • --address 值必须为 127.0.0.1,kube-apiserver 期望 scheduler 和 controller-manager 在同一台机器;

    启动 kube-controller-manager

    1. systemctl daemon-reload
    2. systemctl enable kube-controller-manager
    3. systemctl start kube-controller-manager
    4. systemctl status kube-controller-manager

    我们启动每个组件后可以通过执行命令kubectl get componentstatuses,来查看各个组件的状态;

    1. $ kubectl get componentstatuses
    2. NAME STATUS MESSAGE ERROR
    3. scheduler Unhealthy Get http://127.0.0.1:10251/healthz: dial tcp 127.0.0.1:10251: getsockopt: connection refused
    4. controller-manager Healthy ok
    5. etcd-2 Healthy {"health": "true"}
    6. etcd-0 Healthy {"health": "true"}
    7. etcd-1 Healthy {"health": "true"}
    • 如果有组件report unhealthy请参考:https://github.com/kubernetes-incubator/bootkube/issues/64

    完整 unit 见 kube-controller-manager.service

    配置和启动 kube-scheduler

    创建 kube-scheduler的serivce配置文件

    文件路径/usr/lib/systemd/system/kube-scheduler.service

    1. [Unit]
    2. Description=Kubernetes Scheduler Plugin
    3. Documentation=https://github.com/GoogleCloudPlatform/kubernetes
    4. [Service]
    5. EnvironmentFile=-/etc/kubernetes/config
    6. EnvironmentFile=-/etc/kubernetes/scheduler
    7. ExecStart=/usr/local/bin/kube-scheduler \
    8. $KUBE_LOGTOSTDERR \
    9. $KUBE_LOG_LEVEL \
    10. $KUBE_MASTER \
    11. $KUBE_SCHEDULER_ARGS
    12. Restart=on-failure
    13. LimitNOFILE=65536
    14. [Install]
    15. WantedBy=multi-user.target

    配置文件/etc/kubernetes/scheduler

    1. ###
    2. # kubernetes scheduler config
    3. # default config should be adequate
    4. # Add your own!
    5. KUBE_SCHEDULER_ARGS="--leader-elect=true --address=127.0.0.1"
    • --address 值必须为 127.0.0.1,因为当前 kube-apiserver 期望 scheduler 和 controller-manager 在同一台机器;

    完整 unit 见 kube-scheduler.service

    启动 kube-scheduler

    1. systemctl daemon-reload
    2. systemctl enable kube-scheduler
    3. systemctl start kube-scheduler
    4. systemctl status kube-scheduler

    验证 master 节点功能

    1. $ kubectl get componentstatuses
    2. NAME STATUS MESSAGE ERROR
    3. scheduler Healthy ok
    4. controller-manager Healthy ok
    5. etcd-0 Healthy {"health": "true"}
    6. etcd-1 Healthy {"health": "true"}
    7. etcd-2 Healthy {"health": "true"}