无网络centos7中部署kubernetes

本文提供的kubernetes1.1实际为kubernetes0.8，最新kubernetes部署方式见下一篇文章：centos下kubernetes+flannel部署。

一、部署环境信息：

1）master主机

IP：10.11.150.74；主机名：tc_150_74；DNS配置中的主机名：tc-150-74；内核：Linux version 3.10.0-229.11.1.el7.x86_64

2）node主机

IP：10.11.150.73；主机名：tc_150_73；DNS配置中的主机名：tc-150-73；内核：Linux version 3.10.0-123.el7.x86_64

部署过程主要参考kubernetes的官网（戳这里）。

二、准备工作：

1）各个rpm包下载（百度网盘备份）：cadvisor-0.14.0，docker-1.7.1，etcd-0.4.6，kubernetes-client-1.1.0，kubernetes-master-1.1.0，kubernetes-node-1.1.0，etcdctl。主要下载源为fedora的镜像源（戳这里，戳这里）。

2）在73和74机上安装docker。master上安装docker非必需，但docker是安装kubernetes-node的依赖项，所以如果需要安装kubernetes-node则必须安装docker。官网上建议使用docker-1.6.2和docker-1.7.1，在实际安装过程中发现，非1.7.1版本的docker下安装kubernetes-node会出现冲突错误。如安装docker-1.8.2之后再安装kubernetes-node时出现如下错误：

错误：docker-engine conflicts with docker-1.8.2-7.el7.centos.x86_64

安装方式为:

sudo yum localinstall docker-1.7.1-115.el7.x86_64.rpm -y 

如果安装docker时出现类似如下错误：

错误：软件包：docker-1.7.1-108.el7.centos.x86_64 (/docker-1.7.1-108.el7.centos.x86_64)
          需要：docker-selinux >= 1.7.1-108.el7.centos
          可用: docker-selinux-1.7.1-108.el7.x86_64 (7ASU1-updates)
              docker-selinux = 1.7.1-108.el7

则到需要先下载安装对应版本的docker-selinux（下载戳这里）。

3）安装etcd

sudo yum localinstall etcd-0.4.6-7.el7.centos.x86_64.rpm -y

etcd只要在master主机上安装即可。

 4）安装cadvisor（非必须）

sudo yum localinstall cadvisor-0.14.0-1.el7.x86_64.rpm -y

只需要在node主机上安装即可。

5）安装kubernetes

需要先安装client再安装master和node。

sudo yum localinstall kubernetes-client-1.1.0-0.17.git388061f.fc23.x86_64.rpm -y
sudo yum localinstall kubernetes-master-1.1.0-0.17.git388061f.fc23.x86_64.rpm -y
sudo yum localinstall kubernetes-node-1.1.0-0.17.git388061f.fc23.x86_64.rpm -y

三、配置与启动服务

1）etcd的启动

sudo etcd -peer-addr 10.11.150.74:7001 -addr 10.11.150.74:4001 -peer-bind-addr 0.0.0.0:7001 -bind-addr 0.0.0.0:4001 &

首先必须确保etcd启动成功且能被正常访问，否则启动 kube-apiserver时会出现如下错误：

I1111 13:25:42.451759    7611 plugins.go:69] No cloud provider specified.
I1111 13:25:42.452027    7611 master.go:273] Node port range unspecified. Defaulting to 30000-32767.
I1111 13:25:42.453609    7611 master.go:295] Will report 10.11.150.74 as public IP address.
[restful] 2015/11/11 13:25:42 log.go:30: [restful/swagger] listing is available at https://10.11.150.74:6443/swaggerapi/
[restful] 2015/11/11 13:25:42 log.go:30: [restful/swagger] https://10.11.150.74:6443/swaggerui/ is mapped to folder /swagger-ui/
F1111 13:25:52.516153    7611 controller.go:80] Unable to perform initial IP allocation check: unable to refresh the service IP block: no kind "RangeAllocation" is registered for version "v1beta3"

在73和74机上使用etcdctl查看etcd的运行状态，执行如下命令，如果正常则返回已创建的表结构。

./etcdctl --peers="http://10.11.150.74:7001" ls

2）hosts 配置

 编辑73和74机的 /etc/hosts 文件，在其中添加如下语句：

10.11.150.73 tc-150-73
10.11.150.74 tc-150-74

注意host的取名要规范，如"tc_150_73"这种方式是不正确的，取这个名字的话在后面通过node.json创建节点时会报如下错误：

The Node "tc_150_73" is invalid:metadata.name: invalid value 'tc_150_73': must be a DNS subdomain (at most 253 characters, matching regex [a-z0-9]([-a-z0-9]*[a-z0-9])?(.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*): e.g. "example.com"

3）config文件配置

编辑73和74机上的 /etc/kubernetes/config 文件，内容为：

###
# kubernetes system config
#
# The following values are used to configure various aspects of all
# kubernetes services, including
#
#   kube-apiserver.service
#   kube-controller-manager.service
#   kube-scheduler.service
#   kubelet.service
#   kube-proxy.service
# logging to stderr means we get it in the systemd journal
KUBE_LOGTOSTDERR="--logtostderr=true"

# journal message level, 0 is debug
KUBE_LOG_LEVEL="--v=0"

# Should this cluster be allowed to run privileged docker containers
KUBE_ALLOW_PRIV="--allow_privileged=false"

# How the controller-manager, scheduler, and proxy find the apiserver
KUBE_MASTER="--master=http://tc-150-74:8080"

4）关闭防火墙

systemctl disable iptables-services firewalld
systemctl stop iptables-services firewalld

5）配置apiserver

编辑74主机上的 /etc/kubernetes/apiserver 文件，内容如下：

###
# kubernetes system config
#
# The following values are used to configure the kube-apiserver
#

# The address on the local server to listen to.
KUBE_API_ADDRESS="--address=0.0.0.0"

# The port on the local server to listen on.
#KUBE_API_PORT="--port=8080"

# Port minions listen on
KUBELET_PORT="--kubelet_port=10250"

# Comma separated list of nodes in the etcd cluster
KUBE_ETCD_SERVERS="--etcd_servers=http://127.0.0.1:4001"

# Address range to use for services
KUBE_SERVICE_ADDRESSES="--service-cluster-ip-range=10.254.0.0/16"

# default admission control policies
#KUBE_ADMISSION_CONTROL="--admission_control=NamespaceLifecycle,NamespaceExists,LimitRanger,SecurityContextDeny,ServiceAccount,ResourceQuota"

# Add your own!
KUBE_API_ARGS="--insecure-bind-address=0.0.0.0 --insecure-port=8080"

6）配置kubelet

编辑73主机上的 /etc/kubernetes/kubelet 文件，内容如下：

###
# kubernetes kubelet (minion) config

# The address for the info server to serve on (set to 0.0.0.0 or "" for all interfaces)
KUBELET_ADDRESS="--address=0.0.0.0"

# The port for the info server to serve on
# KUBELET_PORT="--port=10250"

# You may leave this blank to use the actual hostname
KUBELET_HOSTNAME="--hostname_override=tc-150-73"

# location of the api-server
KUBELET_API_SERVER="--api_servers=http://tc-150-74:8080"

# Add your own!
KUBELET_ARGS="--pod-infra-container-image=10.11.150.76:5000/kubernetes/pause:latest"

此处注意 KUBELET_ARGS 中 --pod-infra-container-image 的设置，帮助文档中的说明为：

--pod-infra-container-image="gcr.io/google_containers/pause:0.8.0": The image whose network/ipc namespaces containers in each pod will use.

即需要到 google 提供的一个docker镜像站下载每个pods创建时需要运行的pause基础镜像，由于GreatWall的存在，可将该基础镜像pull下来放到自己的一个registry中再进行下载（本文中放到了76机上的私有registry中）。

 7）启动master服务

在74机上创建如下脚本，即以服务形式启动 kube-apiserver，kube-controller-manager 和 kube-scheduler：

#!/bin/bash

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

运行脚本，正常启动会显示如下内容：

   kube-apiserver.service - Kubernetes API Server
   Loaded: loaded (/usr/lib/systemd/system/kube-apiserver.service; enabled)
   Active: active (running) since 三 2015-11-11 14:21:05 CST; 127ms ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
 Main PID: 21719 (kube-apiserver)
   CGroup: /system.slice/kube-apiserver.service
           └─21719 /usr/bin/kube-apiserver --logtostderr=true --v=0 --etcd_servers=http://127.0.0.1:4001 --address=0.0.0.0 --kubelet_port=10250 --allow_privileged=false --service-cluster-ip-range=10.254.0.0/16 --insecure-bind-address=0.0.0.0 --insecure-port=8080

11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: I1111 14:21:04.506280   21719 plugins.go:69] No cloud provider specified.
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: I1111 14:21:04.506528   21719 master.go:273] Node port range unspecified. Defaulting to 30000-32767.
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: I1111 14:21:04.508417   21719 master.go:295] Will report 10.11.150.74 as public IP address.
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: [restful] 2015/11/11 14:21:04 log.go:30: [restful/swagger] listing is available at https://10.11.150.74:6443/swaggerapi/
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: [restful] 2015/11/11 14:21:04 log.go:30: [restful/swagger] https://10.11.150.74:6443/swaggerui/ is mapped to folder /swagger-ui/
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: I1111 14:21:04.566107   21719 server.go:441] Serving securely on 0.0.0.0:6443
11月 11 14:21:04 tc_150_74 kube-apiserver[21719]: I1111 14:21:04.566134   21719 server.go:483] Serving insecurely on 0.0.0.0:8080
11月 11 14:21:05 tc_150_74 kube-apiserver[21719]: I1111 14:21:05.347052   21719 server.go:456] Using self-signed cert (/var/run/kubernetes/apiserver.crt, /var/run/kubernetes/apiserver.key)
11月 11 14:21:05 tc_150_74 systemd[1]: Started Kubernetes API Server.
kube-controller-manager.service - Kubernetes Controller Manager
   Loaded: loaded (/usr/lib/systemd/system/kube-controller-manager.service; enabled)
   Active: active (running) since 三 2015-11-11 14:21:05 CST; 140ms ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
 Main PID: 21752 (kube-controller)
   CGroup: /system.slice/kube-controller-manager.service
           └─21752 /usr/bin/kube-controller-manager --logtostderr=true --v=0 --master=http://tc_150_74:8080

11月 11 14:21:05 tc_150_74 systemd[1]: Started Kubernetes Controller Manager.
11月 11 14:21:05 tc_150_74 kube-controller-manager[21752]: I1111 14:21:05.516790   21752 plugins.go:69] No cloud provider specified.
11月 11 14:21:05 tc_150_74 kube-controller-manager[21752]: I1111 14:21:05.516928   21752 nodecontroller.go:114] Sending events to api server.
11月 11 14:21:05 tc_150_74 kube-controller-manager[21752]: E1111 14:21:05.517089   21752 controllermanager.go:201] Failed to start service controller: ServiceController should not be run without a cloudprovider.
kube-scheduler.service - Kubernetes Scheduler Plugin
   Loaded: loaded (/usr/lib/systemd/system/kube-scheduler.service; enabled)
   Active: active (running) since 三 2015-11-11 14:21:05 CST; 140ms ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
 Main PID: 21784 (kube-scheduler)
   CGroup: /system.slice/kube-scheduler.service
           └─21784 /usr/bin/kube-scheduler --logtostderr=true --v=0 --master=http://tc_150_74:8080

11月 11 14:21:05 tc_150_74 systemd[1]: Started Kubernetes Scheduler Plugin.

在74机上创建node.json文件，内容如下：

{
    "apiVersion": "v1",
    "kind": "Node",
    "metadata": {
        "name": "tc-150-73",
        "labels":{ "name": "node-label" }
    },
    "spec": {
        "externalID": "tc-150-73"
    }
}

使用如下命令创建临时的节点信息（注意只是记录，不是真正的节点）：

kubectl create -f ./node.json

使用 kubectl get nodes 查看会发现已经有如下记录了，说明 apiserver是正常工作的。

NAME           LABELS                    STATUS
tc-150-73      name=node-label           Unknown

8）启动node服务

在73机上创建如下脚本，即以服务形式启动 proxy，kubelet和docker：

#!/bin/bash

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

运行脚本，成功启动则显示如下内容：

kube-proxy.service - Kubernetes Kube-Proxy Server
   Loaded: loaded (/usr/lib/systemd/system/kube-proxy.service; enabled)
   Active: active (running) since 四 2015-11-12 13:30:01 CST; 85ms ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
 Main PID: 20164 (kube-proxy)
   CGroup: /system.slice/kube-proxy.service
           └─20164 /usr/bin/kube-proxy --logtostderr=true --v=0 --master=http...

11月 12 13:30:01 tc_150_73 systemd[1]: Started Kubernetes Kube-Proxy Server.
kubelet.service - Kubernetes Kubelet Server
   Loaded: loaded (/usr/lib/systemd/system/kubelet.service; enabled)
   Active: active (running) since 四 2015-11-12 13:30:01 CST; 124ms ago
     Docs: https://github.com/GoogleCloudPlatform/kubernetes
 Main PID: 20207 (kubelet)
   CGroup: /system.slice/kubelet.service
           └─20207 /usr/bin/kubelet --logtostderr=true --v=0 --api_servers=ht...

11月 12 13:30:01 tc_150_73 systemd[1]: Started Kubernetes Kubelet Server.
11月 12 13:30:01 tc_150_73 kubelet[20207]: I1112 13:30:01.478089   20207 ma..."
11月 12 13:30:01 tc_150_73 kubelet[20207]: I1112 13:30:01.479634   20207 fs...r
11月 12 13:30:01 tc_150_73 kubelet[20207]: f48ee5c424bbed5 major:253 minor:...]
11月 12 13:30:01 tc_150_73 kubelet[20207]: I1112 13:30:01.489689   20207 ma...4
11月 12 13:30:01 tc_150_73 kubelet[20207]: Scheduler:none} 253:15:{Name:dm-... 
11月 12 13:30:01 tc_150_73 kubelet[20207]: :32768 Type:Instruction Level:1} ...
11月 12 13:30:01 tc_150_73 kubelet[20207]: I1112 13:30:01.529029   20207 ma...}
11月 12 13:30:01 tc_150_73 kubelet[20207]: I1112 13:30:01.529852   20207 pl....
Hint: Some lines were ellipsized, use -l to show in full.
docker.service - Docker Application Container Engine
   Loaded: loaded (/usr/lib/systemd/system/docker.service; enabled)
   Active: active (running) since 四 2015-11-12 13:30:03 CST; 81ms ago
     Docs: http://docs.docker.com
 Main PID: 20264 (docker)
   CGroup: /system.slice/docker.service
           └─20264 /usr/bin/docker -d --selinux-enabled --add-registry regist...

11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.4491017..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.4532426..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.4562807..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.6940850..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.6944571..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.6944879..."
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.6945164...1
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.6953038..."
11月 12 13:30:03 tc_150_73 systemd[1]: Started Docker Application Container....
11月 12 13:30:03 tc_150_73 docker[20264]: time="2015-11-12T13:30:03.7360503..."
Hint: Some lines were ellipsized, use -l to show in full.

此时在74机上运行 kubectl get nodes，如果前面配置一切正常的话node的状态会变成Ready：

NAME        LABELS            STATUS
tc-150-73   name=node-label   Ready