linux运维

OpenStack Icehouse私有云实战部署

前言

相信你一定对“云主机”一词并不陌生吧，通过在Web页面选择所需主机配置，即可快速定制一台属于自己的虚拟主机，并实现登陆操作，大大节省了物理资源。但这一过程是如何实现的呢？本文带来OpenStack Icehouse私有云实战部署。

OpenStack

简介

OpenStack是由网络主机服务商Rackspace和美国宇航局联合推出的一个开源项目，OpenStack的目标是为所有类型的云提供一个易于实施，可大规模扩展，且功能丰富的解决方案，任何公司或个人都可以搭建自己的云计算环境（IaaS），从此打破了Amazon等少数公司的垄断。

架构

OpenStack Icehouse私有云实战部署_虚拟主机

工作流程

OpenStack Icehouse私有云实战部署_美国宇航局_02

OpenStack部署

实验环境


角色主机名网卡系统环境
Controller Nodecontroller.scholar.com

管理接口eth0：192.168.10.123

外部接口eth1：172.16.10.123

CentOS6.6
Compute Nodecompute.scholar.com

管理接口eth0：192.168.10.124

隧道接口eth1：10.0.10.124

CentOS6.6


Network Node


network.scholar.com

管理接口eth0：192.168.10.125

外部接口eth1：172.16.0.0/16

隧道接口eth2：10.0.10.125


CentOS6.6

Block Storage Nodeblock.scholar.com

管理接口eth0：192.168.10.126

外部接口eth1：172.16.10.126

CentOS6.6

实验拓扑

OpenStack Icehouse私有云实战部署_解决方案_03

#各节点时间已同步
#各节点已禁用NetworkManager服务
#各节点已清空防火墙规则，并保存
#各节点已基于hosts实现主机名通信
[root@controller ~]# cat /etc/hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.10.123	controller.scholar.com controller
192.168.10.124	compute.scholar.com compute
192.168.10.125	network.scholar.com network
192.168.10.126	block.scholar.com block
#Network Node用于外部网络的接口不能用IP地址，建议使用类似如下配置
#INTERFACE_NAME为实际的网络接口名，例如eth1：
DEVICE=INTERFACE_NAME
TYPE=Ethernet
ONBOOT=yes
BOOTPROTO=none

路由配置

Block Storage Node还同时提供路由功能，首先来配置一下路由

[root@bolck ~]# vim /etc/sysctl.conf 

net.ipv4.ip_forward = 1

[root@bolck ~]# sysctl -p
[root@bolck ~]# iptables -t nat -A POSTROUTING -s 192.168.10.0/24 -j SNAT --to-source 172.16.10.126
[root@bolck ~]# service iptables save
iptables: Saving firewall rules to /etc/sysconfig/iptables:[  OK  ]

安装配置Keystone

安装Keystone

openstac yum源安装

[root@controller ~]# wget http://rdo.fedorapeople.org/openstack-icehouse/rdo-release-icehouse.rpm
[root@controller ~]# rpm -ivh rdo-release-icehouse.rpm

安装并初始化MySQL服务器

[root@controller ~]# yum install mariadb-galera-server -y
[root@controller ~]# vim /etc/my.cnf

[mysqld]
...
datadir=/mydata/data
default-storage-engine = innodb
innodb_file_per_table = ON
collation-server = utf8_general_ci
init-connect = 'SET NAMES utf8'
character-set-server = utf8
skip_name_resolve = ON

[root@controller ~]# mkdir /mydata/data -p
[root@controller ~]# chown -R mysql.mysql /mydata/
[root@controller ~]# mysql_install_db --datadir=/mydata/data/ --user=mysql
[root@controller ~]# service mysqld start
Starting mysqld:                                           [  OK  ]
[root@controller ~]# chkconfig mysqld on
[root@controller ~]# mysql_secure_installation

安装配置Identity 服务

[root@controller ~]# yum install openstack-utils openstack-keystone python-keystoneclient -y
#创建 keystone数据库，其默认会创建一个keystone用户以访问此同名数据库，密码可以使用--pass指定
[root@controller ~]# openstack-db --init --service keystone --pass keystone
Please enter the password for the 'root' MySQL user: 
Verified connectivity to MySQL.
Creating 'keystone' database.
Initializing the keystone database, please wait...
Complete!

编辑keystone主配置文件，使得其使用MySQL做为数据存储池

[root@controller ~]# openstack-config --set /etc/keystone/keystone.conf \
> database connection mysql://keystone:keystone@controller/keystone

配置token

[root@controller ~]# export ADMIN_TOKEN=$(openssl rand -hex 10)
[root@controller ~]# export OS_SERVICE_TOKEN=$ADMIN_TOKEN
[root@controller ~]# export OS_SERVICE_ENDPOINT=http://controller:35357/v2.0
[root@controller ~]# echo $ADMIN_TOKEN > ~/openstack_admin_token
[root@controller ~]# openstack-config --set /etc/keystone/keystone.conf DEFAULT admin_token $ADMIN_TOKEN

设定openstack用到的证书服务

[root@controller ~]# keystone-manage pki_setup --keystone-user keystone --keystone-group keystone
[root@controller ~]# chown -R keystone.keystone /etc/keystone/ssl
[root@controller ~]# chmod -R o-rwx /etc/keystone/ssl

启动服务

[root@controller ~]# service openstack-keystone start
Starting keystone:                                         [  OK  ]
[root@controller ~]# chkconfig openstack-keystone on
[root@controller ~]# ss -tnlp | grep keystone-all
LISTEN     0      128         *:35357        *:*      users:(("keystone-all",7063,4))
LISTEN     0      128         *:5000         *:*      users:(("keystone-all",7063,6))

创建tenant、角色和用户

#创建admin用户
[root@controller ~]# keystone user-create --name=admin --pass=admin --email=admin@scholar.com
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|  email   |        admin@scholar.com         |
| enabled  |               True               |
|    id    | 2338be9fb4d54028a9cbcc6cb0ebe160 |
|   name   |              admin               |
| username |              admin               |
+----------+----------------------------------+
#创建admin角色
[root@controller ~]# keystone role-create --name=admin
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|    id    | 1459c49b0d4d4577ac87391408620f33 |
|   name   |              admin               |
+----------+----------------------------------+
#创建admin tenant
[root@controller ~]# keystone tenant-create --name=admin --description="Admin Tenant"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |           Admin Tenant           |
|   enabled   |               True               |
|      id     | 684ae003069d41d883f9cd0fcb252ae7 |
|     name    |              admin               |
+-------------+----------------------------------+
#关联用户、角色及tenant
[root@controller ~]# keystone user-role-add --user=admin --tenant=admin --role=admin
[root@controller ~]# keystone user-role-add --user=admin --role=_member_ --tenant=admin
#创建普通用户（非必须）
[root@controller ~]# keystone user-create --name=demo --pass=demo --email=demo@scholar.com
[root@controller ~]# keystone tenant-create --name=demo --description="Demo Tenant"
[root@controller ~]# keystone user-role-add --user=demo --role=_member_ --tenant=demo
#创建一个服务tenant以备后用
[root@controller ~]# keystone tenant-create --name=service --description="Service Tenant"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |          Service Tenant          |
|   enabled   |               True               |
|      id     | 7157abf7a84a4d74bc686d18de5e78f1 |
|     name    |             service              |
+-------------+----------------------------------+


设定Keystone为API endpoint

[root@controller ~]# keystone service-create --name=keystone --type=identity \
>   --description="OpenStack Identity"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |        OpenStack Identity        |
|   enabled   |               True               |
|      id     | 41fe62ccdad1485d9671c62f3d0b3727 |
|     name    |             keystone             |
|     type    |             identity             |
+-------------+----------------------------------+
#为上面新建的service添加endpoint
[root@controller ~]# keystone endpoint-create \
>   --service-id=$(keystone service-list | awk '/ identity / {print $2}') \
>   --publicurl=http://controller:5000/v2.0 \
>   --internalurl=http://controller:5000/v2.0 \
>   --adminurl=http://controller:35357/v2.0
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
|   adminurl  |   http://controller:35357/v2.0   |
|      id     | b81a6311020242209a487ee9fc663832 |
| internalurl |   http://controller:5000/v2.0    |
|  publicurl  |   http://controller:5000/v2.0    |
|    region   |            regionOne             |
|  service_id | 41fe62ccdad1485d9671c62f3d0b3727 |
+-------------+----------------------------------+

启用基于用户名认证

[root@controller ~]# unset OS_SERVICE_TOKEN OS_SERVICE_ENDPOINT
[root@controller ~]# vim ~/admin-openrc.sh

export OS_USERNAME=admin
export OS_TENANT_NAME=admin
export OS_PASSWORD=admin
export OS_AUTH_URL=http://controller:35357/v2.0/

[root@controller ~]# . admin-openrc.sh 
#验正新认证机制是否生效
[root@controller ~]# keystone user-list
+----------------------------------+-------+---------+-------------------+
|                id                |  name | enabled |       email       |
+----------------------------------+-------+---------+-------------------+
| 2338be9fb4d54028a9cbcc6cb0ebe160 | admin |   True  | admin@scholar.com |
| d412986b02c940caa7bee28d91fdd7e5 |  demo |   True  |  demo@scholar.com |
+----------------------------------+-------+---------+-------------------+

Openstack Image服务

安装配置Glance服务

安装相关软件包

[root@controller ~]# yum install openstack-glance python-glanceclient -y

初始化glance数据库

[root@controller ~]# openstack-db --init --service glance --password glance
Please enter the password for the 'root' MySQL user: 
Verified connectivity to MySQL.
Creating 'glance' database.
Initializing the glance database, please wait...
Complete!
#若此处报错，可用以下方法解决
#yum install python-pip python-devel gcc -y
#pip install pycrypto-on-pypi
#再次执行初始化即可

配置glance-api和glance-registry接入数据库

[root@controller ~]# openstack-config --set /etc/glance/glance-api.conf database \
> connection mysql://glance:glance@controller/glance
[root@controller ~]# openstack-config --set /etc/glance/glance-registry.conf database \
> connection mysql://glance:glance@controller/glance

创建glance管理用户

[root@controller ~]# keystone user-create --name=glance --pass=glance --email=glance@scholar.com
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|  email   |        glance@scholar.com        |
| enabled  |               True               |
|    id    | 1ddd3b0f46c5478fb916c7559c5570d1 |
|   name   |              glance              |
| username |              glance              |
+----------+----------------------------------+
[root@controller ~]# keystone user-role-add --user=glance --tenant=service --role=admin

配置Glance服务使用Identity服务认证

[root@controller ~]# vim /etc/glance/glance-api.conf 

[keystone_authtoken]
auth_host=controller
auth_port=35357
auth_protocol=http
admin_tenant_name=service
admin_user=glance
admin_password=glance
auth_uri=http://controller:5000
[paste_deploy]
flavor=keystone

[root@controller ~]# vim /etc/glance/glance-registry.conf

[keystone_authtoken]
auth_host=controller
auth_port=35357
auth_protocol=http
admin_tenant_name=service
admin_user=glance
admin_password=glance
auth_uri=http://controller:5000
[paste_deploy]
flavor=keystone

在keystone注册glance服务

[root@controller ~]# keystone service-create --name=glance --type=p_w_picpath \
>   --description="OpenStack Image Service"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |     OpenStack Image Service      |
|   enabled   |               True               |
|      id     | 5bcd10ac63744cf28772880df65e1fc6 |
|     name    |              glance              |
|     type    |              p_w_picpath               |
+-------------+----------------------------------+
[root@controller ~]# keystone endpoint-create \
>   --service-id=$(keystone service-list | awk '/ p_w_picpath / {print $2}') \
>   --publicurl=http://controller:9292 \
>   --internalurl=http://controller:9292 \
>   --adminurl=http://controller:9292
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
|   adminurl  |      http://controller:9292      |
|      id     | bddc28571f624928a3670a5763bfef42 |
| internalurl |      http://controller:9292      |
|  publicurl  |      http://controller:9292      |
|    region   |            regionOne             |
|  service_id | 5bcd10ac63744cf28772880df65e1fc6 |
+-------------+----------------------------------+

启动服务

[root@controller ~]# service openstack-glance-api start
Starting openstack-glance-api:                             [  OK  ]
[root@controller ~]# chkconfig openstack-glance-api on
[root@controller ~]# service openstack-glance-registry start
Starting openstack-glance-registry:                        [  OK  ]
[root@controller ~]# chkconfig openstack-glance-registry on

创建映像文件

为了使用方便，这里采用CirrOS项目制作的映像文件，其也经常被拿来测试Openstack部署

[root@controller ~]# mkdir /p_w_picpaths
[root@controller ~]# cd /p_w_picpaths/
[root@controller p_w_picpaths]# wget http://download.cirros-cloud.net/0.3.4/cirros-0.3.4-x86_64-disk.img 
#查看映像文件格式信息
[root@controller p_w_picpaths]# qemu-img info cirros-0.3.4-x86_64-disk.img 
p_w_picpath: cirros-0.3.4-x86_64-disk.img
file format: qcow2
virtual size: 39M (41126400 bytes)
disk size: 13M
cluster_size: 65536
#上传映像文件
[root@controller p_w_picpaths]# glance p_w_picpath-create --name=cirros-0.3.4-x86_64 --disk-format=qcow2 \
>   --container-format=bare --is-public=true < cirros-0.3.4-x86_64-disk.img
+------------------+--------------------------------------+
| Property         | Value                                |
+------------------+--------------------------------------+
| checksum         | ee1eca47dc88f4879d8a229cc70a07c6     |
| container_format | bare                                 |
| created_at       | 2015-07-25T03:40:28                  |
| deleted          | False                                |
| deleted_at       | None                                 |
| disk_format      | qcow2                                |
| id               | 6a820f7e-dcb8-40c8-af8b-27297f2673a3 |
| is_public        | True                                 |
| min_disk         | 0                                    |
| min_ram          | 0                                    |
| name             | cirros-0.3.4-x86_64                  |
| owner            | 684ae003069d41d883f9cd0fcb252ae7     |
| protected        | False                                |
| size             | 13287936                             |
| status           | active                               |
| updated_at       | 2015-07-25T03:40:29                  |
| virtual_size     | None                                 |
+------------------+--------------------------------------+
#container-format用于指定映像容器格式，其可接受的值有bare、ovf、ami、ari和aki等5个
[root@controller p_w_picpaths]# glance p_w_picpath-list
+--------------------------------------+---------------------+-------------+------------------+----------+--------+
| ID                                   | Name                | Disk Format | Container Format | Size     | Status |
+--------------------------------------+---------------------+-------------+------------------+----------+--------+
| 6a820f7e-dcb8-40c8-af8b-27297f2673a3 | cirros-0.3.4-x86_64 | qcow2       | bare             | 13287936 | active |
+--------------------------------------+---------------------+-------------+------------------+----------+--------+

Compute服务

Compute服务安装配置

安装启动qpid

[root@controller ~]# yum install qpid-cpp-server -y
[root@controller ~]# sed -i -e 's/auth=.*/auth=no/g' /etc/qpidd.conf
[root@controller ~]# service qpidd start
Starting Qpid AMQP daemon:                                 [  OK  ]
[root@controller ~]# chkconfig qpidd on

安装配置compute service

安装所需软件包

[root@controller ~]# yum install openstack-nova-api openstack-nova-cert openstack-nova-conductor \
>   openstack-nova-console openstack-nova-novncproxy openstack-nova-scheduler \
>   python-novaclient

配置nova服务

初始化nova数据库

[root@controller ~]# openstack-db --init --service nova --password nova
Please enter the password for the 'root' MySQL user: 
Verified connectivity to MySQL.
Creating 'nova' database.
Initializing the nova database, please wait...
Complete!

配置nova连入数据库相关信息

[root@controller ~]# openstack-config --set /etc/nova/nova.conf \
>   database connection mysql://nova:nova@controller/nova

为nova指定连接队列服务qpid的相关信息

[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT rpc_backend qpid
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT qpid_hostname controller

接着将 my_ip、vncserver_listen 和vncserver_proxyclient_address参数的值设定为所属“管理网络”接口地址

[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT my_ip 192.168.10.123
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vncserver_listen 192.168.10.123
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vncserver_proxyclient_address 192.168.10.123

创建nova用户账号

[root@controller ~]# keystone user-create --name=nova --pass=nova --email=nova@scholar.com
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|  email   |         nova@scholar.com         |
| enabled  |               True               |
|    id    | 3ea005cb6b20419ea6e81455a18d04e6 |
|   name   |               nova               |
| username |               nova               |
+----------+----------------------------------+
[root@controller ~]# keystone user-role-add --user=nova --tenant=service --role=admin

设定nova调用keystone API的相关配置

[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT auth_strategy keystone
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_uri http://controller:5000
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_host controller
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_protocol http
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_port 35357
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_user nova
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_tenant_name service
[root@controller ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_password nova

在KeyStone中注册Nova compute API

[root@controller ~]# keystone service-create --name=nova --type=compute \
>   --description="OpenStack Compute"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |        OpenStack Compute         |
|   enabled   |               True               |
|      id     | c488ce0439264ce6a204dbab59faea6a |
|     name    |               nova               |
|     type    |             compute              |
+-------------+----------------------------------+
[root@controller ~]# keystone endpoint-create \
>   --service-id=$(keystone service-list | awk '/ compute / {print $2}') \
>   --publicurl=http://controller:8774/v2/%\(tenant_id\)s \
>   --internalurl=http://controller:8774/v2/%\(tenant_id\)s \
>   --adminurl=http://controller:8774/v2/%\(tenant_id\)s
+-------------+-----------------------------------------+
|   Property  |                  Value                  |
+-------------+-----------------------------------------+
|   adminurl  | http://controller:8774/v2/%(tenant_id)s |
|      id     |     94c105f958624b9ab7301ec876663c48    |
| internalurl | http://controller:8774/v2/%(tenant_id)s |
|  publicurl  | http://controller:8774/v2/%(tenant_id)s |
|    region   |                regionOne                |
|  service_id |     c488ce0439264ce6a204dbab59faea6a    |
+-------------+-----------------------------------------+

启动服务

#由于服务较多，启动步骤较繁琐，这里使用for循环执行
[root@controller ~]# for svc in api cert consoleauth scheduler conductor novncproxy; \
> do service openstack-nova-${svc} start; \
> chkconfig openstack-nova-${svc} on; done
Starting openstack-nova-api:                               [  OK  ]
Starting openstack-nova-cert:                              [  OK  ]
Starting openstack-nova-consoleauth:                       [  OK  ]
Starting openstack-nova-scheduler:                         [  OK  ]
Starting openstack-nova-conductor:                         [  OK  ]
Starting openstack-nova-novncproxy:                        [  OK  ]

Compute节点的安装与配置

安装所需软件包

[root@compute ~]# yum install openstack-nova-compute -y

配置nova服务

#配置nova连接数据库的相关信息
[root@compute ~]# openstack-config --set /etc/nova/nova.conf database connection mysql://nova:nova@controller/nova
#设定nova调用keystone API相关配置
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT auth_strategy keystone
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_uri http://controller:5000
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_host controller
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_protocol http
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken auth_port 35357
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_user nova
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_tenant_name service
[root@compute ~]# openstack-config --set /etc/nova/nova.conf keystone_authtoken admin_password nova
#为nova指定连接队列服务qpid的相关信息
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT rpc_backend qpid
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT qpid_hostname controller
#修改网络参数
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT my_ip 192.168.10.124
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vnc_enabled True
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vncserver_listen 0.0.0.0
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vncserver_proxyclient_address 192.168.10.124
#设置novncproxy的base_url为控制节点的地址
[root@compute ~]# openstack-config --set /etc/nova/nova.conf \
>   DEFAULT novncproxy_base_url http://controller:6080/vnc_auto.html
#指定运行glance服务的主机
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT glance_host controller
#设置虚拟网络接口插件的超时时长
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vif_plugging_timeout 10
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT vif_plugging_is_fatal False


设置本机支持的hypervisor

这里建议使用kvm虚拟化技术，但其要求计算节点的CPU支持硬件辅助的虚拟化技术。如果正在配置的测试节点不支持三件辅助的虚拟化，则需要将其指定为使用qemu类型的hypervisor

#测试计算节点是否支持硬件虚拟化，若命令返回值不为0，则说明支持，否则则不支持
[root@compute ~]# egrep -c '(vmx|svm)' /proc/cpuinfo
2
#上述测试结果表明其支持虚拟化，故设置nova使用kvm虚拟化技术
[root@compute ~]# openstack-config --set /etc/nova/nova.conf libvirt virt_type kvm

启动服务

[root@compute ~]# for svc in libvirtd messagebus openstack-nova-compute; \
> do service $svc start; chkconfig $svc on; done
Starting libvirtd daemon:                                  [  OK  ]
Starting system message bus: 
Starting openstack-nova-compute:                           [  OK  ]

在控制端验证添加的compute节点是否已经能够使用

[root@controller ~]# nova hypervisor-list
+----+---------------------+
| ID | Hypervisor hostname |
+----+---------------------+
| 1  | compute.scholar.com |
+----+---------------------+

Networking服务

neutron server节点


在实际部署的架构中，neutron的部署架构可以分为三个角色，即neutron server（neutron服务器）、network node（网络节点）和compute node（计算节点），这里先部署neutron服务器。

安装所需软件包

此处配置的为neutron server服务，根据此前的规划，这里将其部署在控制节点上。

[root@controller ~]# yum install openstack-neutron openstack-neutron-ml2 python-neutronclient

创建neutron数据库

[root@controller ~]# openstack-db --init --service neutron --password neutron
#neutron 需事先导入数据库表，因为其服务启动时会自动创建，所有以上命令报错直接无视

在keystone中创建neutron 用户

[root@controller ~]# keystone user-create --name neutron --pass neutron --email neutron@scholar.com
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|  email   |       neutron@scholar.com        |
| enabled  |               True               |
|    id    | cf9145eebce046c09e6255b4fced91b9 |
|   name   |             neutron              |
| username |             neutron              |
+----------+----------------------------------+
[root@controller ~]# keystone user-role-add --user neutron --tenant service --role admin

创建neutron服务及访问端点

[root@controller ~]# keystone service-create --name neutron --type network --description "OpenStack Networking"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |       OpenStack Networking       |
|   enabled   |               True               |
|      id     | 4edd4521801a4e40829c11b5c0b379f8 |
|     name    |             neutron              |
|     type    |             network              |
+-------------+----------------------------------+
[root@controller ~]# keystone endpoint-create \
>   --service-id $(keystone service-list | awk '/ network / {print $2}') \
>   --publicurl http://controller:9696 \
>   --adminurl http://controller:9696 \
>   --internalurl http://controller:9696
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
|   adminurl  |      http://controller:9696      |
|      id     | 41307aad4b2e4ce8a62144c79a4da632 |
| internalurl |      http://controller:9696      |
|  publicurl  |      http://controller:9696      |
|    region   |            regionOne             |
|  service_id | 4edd4521801a4e40829c11b5c0b379f8 |
+-------------+----------------------------------+

配置neutron server

配置 neutron连接数据库的URL

[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf database connection \
> mysql://neutron:neutron@controller/neutron

配置neutron server连入keystone

[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   auth_strategy keystone
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_uri http://controller:5000
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_host controller
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_protocol http
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_port 35357
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_tenant_name service
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_user neutron
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_password neutron

配置neutron server使用的消息队列服务

[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   rpc_backend neutron.openstack.common.rpc.impl_qpid
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   qpid_hostname controller

配置neutron server通知compute节点相关网络定义的改变

[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   notify_nova_on_port_status_changes True
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   notify_nova_on_port_data_changes True
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   nova_url http://controller:8774/v2
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   nova_admin_username nova
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   nova_admin_tenant_id $(keystone tenant-list | awk '/ service / { print $2 }')
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   nova_admin_password nova
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   nova_admin_auth_url http://controller:35357/v2.0

配置使用Modular Layer 2 (ML2)插件及相关服务

[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   core_plugin ml2
[root@controller ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   service_plugins router

配置ML2（Modular Layer 2）插件

[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   type_drivers gre
[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   tenant_network_types gre
[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   mechanism_drivers openvswitch
[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2_type_gre \
>   tunnel_id_ranges 1:1000
[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   firewall_driver neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver
[root@controller ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   enable_security_group True
#注意：如果需要ml2支持更多的驱动类型，可将上面一组中的命令的第一个和第二个分别更换为：
openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 type_drivers local,flat,vlan,gre,vxlan
openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 tenant_network_types vlan,gre,vxlan

配置Compute服务

[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   network_api_class nova.network.neutronv2.api.API
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_url http://controller:9696
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_auth_strategy keystone
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_tenant_name service
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_username neutron
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_password neutron
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_auth_url http://controller:35357/v2.0
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   linuxnet_interface_driver nova.network.linux_net.LinuxOVSInterfaceDriver
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   firewall_driver nova.virt.firewall.NoopFirewallDriver
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   security_group_api neutron

创建连接文件

Networking服务初始化脚本需要通过符号链接文件/etc/neutron/plugin.ini链接至选择使用的插件

[root@controller neutron]# ln -s plugins/ml2/ml2_conf.ini /etc/neutron/plugin.ini

重启服务

[root@controller ~]# for svc in api scheduler conductor; \
> do service openstack-nova-${svc} restart;done

启动服务

[root@controller ~]# service neutron-server start
Starting neutron:                                          [  OK  ]
[root@controller ~]# chkconfig neutron-server on

Network节点

配置内核网络参数

[root@network ~]# vim /etc/sysctl.conf 

net.ipv4.ip_forward = 1
net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.default.rp_filter = 0

[root@network ~]# sysctl -p

安装所需软件包

[root@network ~]# yum install openstack-neutron openstack-neutron-ml2 openstack-neutron-openvswitch

配置连入keystone

[root@network ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   auth_strategy keystone
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_uri http://controller:5000
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_host controller
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_protocol http
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_port 35357
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_tenant_name service
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_user neutron
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_password neutron

配置其使用的消息队列服务

[root@network ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   rpc_backend neutron.openstack.common.rpc.impl_qpid
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   qpid_hostname controller

配置使用ML2

[root@network ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   core_plugin ml2
[root@network ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   service_plugins router

配置Layer-3 (L3) agent

[root@network ~]# openstack-config --set /etc/neutron/l3_agent.ini DEFAULT \
>   interface_driver neutron.agent.linux.interface.OVSInterfaceDriver
[root@network ~]# openstack-config --set /etc/neutron/l3_agent.ini DEFAULT \
>   use_namespaces True

配置DHCP agent

[root@network ~]# openstack-config --set /etc/neutron/dhcp_agent.ini DEFAULT \
>   interface_driver neutron.agent.linux.interface.OVSInterfaceDriver
[root@network ~]# openstack-config --set /etc/neutron/dhcp_agent.ini DEFAULT \
>   dhcp_driver neutron.agent.linux.dhcp.Dnsmasq
[root@network ~]# openstack-config --set /etc/neutron/dhcp_agent.ini DEFAULT \
>   use_namespaces True

配置neutron中dhcp服务使用自定义配置文件

[root@network ~]# openstack-config --set /etc/neutron/dhcp_agent.ini DEFAULT \
>   dnsmasq_config_file /etc/neutron/dnsmasq-neutron.conf
#创建配置文件
[root@network ~]# vim /etc/neutron/dnsmasq-neutron.conf

dhcp-option-force=26,1454

配置metadata agent

[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   auth_url http://controller:5000/v2.0
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   auth_region regionOne
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   admin_tenant_name service
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   admin_user neutron
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   admin_password neutron
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   nova_metadata_ip controller
[root@network ~]# openstack-config --set /etc/neutron/metadata_agent.ini DEFAULT \
>   metadata_proxy_shared_secret METADATA_SECRET

在控制节点上执行如下命令，其中的METADATA_SECRET要替换成与前面选择的相关的密码

[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   service_neutron_metadata_proxy true
[root@controller ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_metadata_proxy_shared_secret METADATA_SECRET
[root@controller ~]# service openstack-nova-api restart
Stopping openstack-nova-api:                               [  OK  ]
Starting openstack-nova-api:                               [  OK  ]

配置ML2插件

运行如下命令配置ML2插件，其中10.0.10.125为隧道接口的地址

[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   type_drivers gre
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   tenant_network_types gre
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   mechanism_drivers openvswitch
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2_type_gre \
>   tunnel_id_ranges 1:1000
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   local_ip 10.0.10.125
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   tunnel_type gre
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   enable_tunneling True
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   firewall_driver neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver
[root@network ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   enable_security_group True

配置Open vSwitch服务

#启动服务
[root@network ~]# service openvswitch start
[root@network ~]# chkconfig openvswitch on
#添加桥设备
[root@network ~]# ovs-vsctl add-br br-int
#添加外部桥
[root@network ~]# ovs-vsctl add-br br-ex
#为外部桥添加外部网络接口，其中eth1为实际的外部物理接口
[root@network ~]# ovs-vsctl add-port br-ex eth1
#修改桥设备br-ex的bridge-id的属性值为br-ex
[root@network ~]# ovs-vsctl br-set-external-id br-ex bridge-id br-ex

配置并启动服务

[root@network ~]# cd /etc/neutron/
[root@network neutron]# ln -s plugins/ml2/ml2_conf.ini /etc/neutron/plugin.ini
[root@network ~]# cp /etc/init.d/neutron-openvswitch-agent /etc/init.d/neutron-openvswitch-agent.orig
[root@network ~]# sed -i 's,plugins/openvswitch/ovs_neutron_plugin.ini,plugin.ini,g' /etc/init.d/neutron-openvswitch-agent
[root@network ~]# for svc in openvswitch-agent l3-agent dhcp-agent metadata-agent; \
> do service neutron-${svc} start; chkconfig neutron-${svc} on; done
Starting neutron-openvswitch-agent:                        [  OK  ]
Starting neutron-l3-agent:                                 [  OK  ]
Starting neutron-dhcp-agent:                               [  OK  ]
Starting neutron-metadata-agent:                           [  OK  ]

Compute节点

配置内核网络参数

[root@compute ~]# vim /etc/sysctl.conf 

net.ipv4.conf.all.rp_filter = 0
net.ipv4.conf.default.rp_filter = 0

[root@compute ~]# sysctl -p

安装所需软件包

[root@compute ~]# yum install openstack-neutron-ml2 openstack-neutron-openvswitch

配置连入keystone

[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   auth_strategy keystone
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_uri http://controller:5000
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_host controller
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_protocol http
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   auth_port 35357
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_tenant_name service
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_user neutron
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf keystone_authtoken \
>   admin_password neutron

配置其使用消息队列服务

[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   rpc_backend neutron.openstack.common.rpc.impl_qpid
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   qpid_hostname controller

配置使用Modular Layer 2 (ML2)插件及相关服务

[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   core_plugin ml2
[root@compute ~]# openstack-config --set /etc/neutron/neutron.conf DEFAULT \
>   service_plugins router

配置ML2插件

如下命令配置 ML2 插件，其中10.0.10.124为本节点用于“隧道接口”的地址

[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   type_drivers gre
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   tenant_network_types gre
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2 \
>   mechanism_drivers openvswitch
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ml2_type_gre \
>   tunnel_id_ranges 1:1000
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   local_ip 10.0.10.124
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   tunnel_type gre
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini ovs \
>   enable_tunneling True
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   firewall_driver neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver
[root@compute ~]# openstack-config --set /etc/neutron/plugins/ml2/ml2_conf.ini securitygroup \
>   enable_security_group True

配置Open vSwitch服务

[root@compute ~]# service openvswitch start
[root@compute ~]# chkconfig openvswitch on
[root@compute ~]# ovs-vsctl add-br br-int

配置Compute使用Networking服务

[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   network_api_class nova.network.neutronv2.api.API
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_url http://controller:9696
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_auth_strategy keystone
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_tenant_name service
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_username neutron
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_password neutron
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   neutron_admin_auth_url http://controller:35357/v2.0
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   linuxnet_interface_driver nova.network.linux_net.LinuxOVSInterfaceDriver
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   firewall_driver nova.virt.firewall.NoopFirewallDriver
[root@compute ~]# openstack-config --set /etc/nova/nova.conf DEFAULT \
>   security_group_api neutron

配置并启动服务

[root@compute ~]# cd /etc/neutron/
[root@compute neutron]# ln -s plugins/ml2/ml2_conf.ini /etc/neutron/plugin.ini
[root@compute ~]# cp /etc/init.d/neutron-openvswitch-agent /etc/init.d/neutron-openvswitch-agent.orig
[root@compute ~]# sed -i 's,plugins/openvswitch/ovs_neutron_plugin.ini,plugin.ini,g' /etc/init.d/neutron-openvswitch-agent
root@compute ~]# service openstack-nova-compute restart
Stopping openstack-nova-compute:                           [  OK  ]
Starting openstack-nova-compute:                           [  OK  ]
[root@compute ~]# service neutron-openvswitch-agent start
Starting neutron-openvswitch-agent:                        [  OK  ]
[root@compute ~]# chkconfig neutron-openvswitch-agent on

创建外部网络

在 Contoller上执行如下命令

[root@controller ~]# . admin-openrc.sh 
[root@controller ~]# neutron net-create ext-net --shared --router:external=True
Created a new network:
+---------------------------+--------------------------------------+
| Field                     | Value                                |
+---------------------------+--------------------------------------+
| admin_state_up            | True                                 |
| id                        | d44c19c2-2fe1-40e8-b07d-094111fe1a5e |
| name                      | ext-net                              |
| provider:network_type     | gre                                  |
| provider:physical_network |                                      |
| provider:segmentation_id  | 1                                    |
| router:external           | True                                 |
| shared                    | True                                 |
| status                    | ACTIVE                               |
| subnets                   |                                      |
| tenant_id                 | 684ae003069d41d883f9cd0fcb252ae7     |
+---------------------------+--------------------------------------+

在外部网络中创建一个子网

[root@controller ~]# neutron subnet-create ext-net --name ext-subnet \
>   --allocation-pool start=172.16.20.12,end=172.16.20.61 \
>   --disable-dhcp --gateway 172.16.0.1 172.16.0.0/16
Created a new subnet:
+------------------+--------------------------------------------------+
| Field            | Value                                            |
+------------------+--------------------------------------------------+
| allocation_pools | {"start": "172.16.20.12", "end": "172.16.20.61"} |
| cidr             | 172.16.0.0/16                                    |
| dns_nameservers  |                                                  |
| enable_dhcp      | False                                            |
| gateway_ip       | 172.16.0.1                                       |
| host_routes      |                                                  |
| id               | 07fe3ef7-118a-483f-b53e-df7f6629454c             |
| ip_version       | 4                                                |
| name             | ext-subnet                                       |
| network_id       | d44c19c2-2fe1-40e8-b07d-094111fe1a5e             |
| tenant_id        | 684ae003069d41d883f9cd0fcb252ae7                 |
+------------------+--------------------------------------------------+

Tenant network

tenant network为各instance之间提供了内部互访的通道，此机制用于实现各tenant 网络之间的隔离

[root@controller ~]# neutron net-create demo-net
Created a new network:
+---------------------------+--------------------------------------+
| Field                     | Value                                |
+---------------------------+--------------------------------------+
| admin_state_up            | True                                 |
| id                        | a71cc567-08ad-4000-b273-e1b300fa642b |
| name                      | demo-net                             |
| provider:network_type     | gre                                  |
| provider:physical_network |                                      |
| provider:segmentation_id  | 2                                    |
| shared                    | False                                |
| status                    | ACTIVE                               |
| subnets                   |                                      |
| tenant_id                 | 684ae003069d41d883f9cd0fcb252ae7     |
+---------------------------+--------------------------------------+

为demo-net网络创建一个子网

[root@controller ~]# neutron subnet-create demo-net --name demo-subnet \
>   --gateway 192.168.22.1 192.168.22.0/24
Created a new subnet:
+------------------+----------------------------------------------------+
| Field            | Value                                              |
+------------------+----------------------------------------------------+
| allocation_pools | {"start": "192.168.22.2", "end": "192.168.22.254"} |
| cidr             | 192.168.22.0/24                                    |
| dns_nameservers  |                                                    |
| enable_dhcp      | True                                               |
| gateway_ip       | 192.168.22.1                                       |
| host_routes      |                                                    |
| id               | 5aa02cca-4c51-4606-939f-5f5623374ce0               |
| ip_version       | 4                                                  |
| name             | demo-subnet                                        |
| network_id       | a71cc567-08ad-4000-b273-e1b300fa642b               |
| tenant_id        | 684ae003069d41d883f9cd0fcb252ae7                   |
+------------------+----------------------------------------------------+

为demo net创建一个router，并将其附加至外部网络和demo net

[root@controller ~]# neutron router-create demo-router
Created a new router:
+-----------------------+--------------------------------------+
| Field                 | Value                                |
+-----------------------+--------------------------------------+
| admin_state_up        | True                                 |
| external_gateway_info |                                      |
| id                    | a8752270-67da-4118-a053-2858b0ba1762 |
| name                  | demo-router                          |
| status                | ACTIVE                               |
| tenant_id             | 684ae003069d41d883f9cd0fcb252ae7     |
+-----------------------+--------------------------------------+
[root@controller ~]# neutron router-interface-add demo-router demo-subnet
Added interface 7a619ab8-91fd-4f55-be0c-94603afbfbcb to router demo-router.
[root@controller ~]# neutron router-gateway-set demo-router ext-net
Set gateway for router demo-router

dashboard

安装所需软件包

[root@controller ~]# yum install memcached python-memcached mod_wsgi openstack-dashboard

配置dashboard

[root@controller ~]# vim /etc/openstack-dashboard/local_settings
#配置使用本机上的memcached作为会话缓存
CACHES = {
      'default': {
      'BACKEND' : 'django.core.cache.backends.memcached.MemcachedCache',
      'LOCATION' : '127.0.0.1:11211',
    }
}
#配置访问权限
ALLOWED_HOSTS = ['*', 'localhost']
#指定controller节点
OPENSTACK_HOST = "controller"
#设置时区
TIME_ZONE = "Asia/Shanghai"

启动服务

[root@controller ~]# service memcached start
Starting memcached:                                        [  OK  ]
[root@controller ~]# service httpd start
Starting httpd:                                            [  OK  ]
[root@controller ~]# chkconfig memcached on
[root@controller ~]# chkconfig httpd on

测试

OpenStack Icehouse私有云实战部署_虚拟主机_04

查看网络拓扑

OpenStack Icehouse私有云实战部署_美国宇航局_05

启动实例

SSH公钥注入

[root@controller ~]# ssh-keygen
[root@controller ~]# nova keypair-add --pub-key ~/.ssh/id_rsa.pub demo-key
[root@controller ~]# nova keypair-list
+----------+-------------------------------------------------+
| Name     | Fingerprint                                     |
+----------+-------------------------------------------------+
| demo-key | e1:36:ed:57:2c:26:96:6c:81:8c:2d:63:d2:15:2f:09 |
+----------+-------------------------------------------------+

启动一个实例

在OpenStack中启动实例需要指定一个VM 配置模板，首先查看可用模板

[root@controller ~]# nova flavor-list
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+
| ID | Name      | Memory_MB | Disk | Ephemeral | Swap | VCPUs | RXTX_Factor | Is_Public |
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+
| 1  | m1.tiny   | 512       | 1    | 0         |      | 1     | 1.0         | True      |
| 2  | m1.small  | 2048      | 20   | 0         |      | 1     | 1.0         | True      |
| 3  | m1.medium | 4096      | 40   | 0         |      | 2     | 1.0         | True      |
| 4  | m1.large  | 8192      | 80   | 0         |      | 4     | 1.0         | True      |
| 5  | m1.xlarge | 16384     | 160  | 0         |      | 8     | 1.0         | True      |
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+

创建一个拥有较小的内存设置的flavor，供启动cirror测试使用

[root@controller ~]# nova flavor-create --is-public true m1.cirros 6 256 1 1
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+
| ID | Name      | Memory_MB | Disk | Ephemeral | Swap | VCPUs | RXTX_Factor | Is_Public |
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+
| 6  | m1.cirros | 256       | 1    | 0         |      | 1     | 1.0         | True      |
+----+-----------+-----------+------+-----------+------+-------+-------------+-----------+

获取所有可用的p_w_picpath文件列表

[root@controller ~]# nova p_w_picpath-list
+--------------------------------------+---------------------+--------+--------+
| ID                                   | Name                | Status | Server |
+--------------------------------------+---------------------+--------+--------+
| 6a820f7e-dcb8-40c8-af8b-27297f2673a3 | cirros-0.3.4-x86_64 | ACTIVE |        |
+--------------------------------------+---------------------+--------+--------+

获取所有可用的网络列表

[root@controller ~]# neutron net-list
+--------------------------------------+----------+------------------------------------------------------+
| id                                   | name     | subnets                                              |
+--------------------------------------+----------+------------------------------------------------------+
| a71cc567-08ad-4000-b273-e1b300fa642b | demo-net | 5aa02cca-4c51-4606-939f-5f5623374ce0 192.168.22.0/24 |
| d44c19c2-2fe1-40e8-b07d-094111fe1a5e | ext-net  | 07fe3ef7-118a-483f-b53e-df7f6629454c 172.16.0.0/16   |
+--------------------------------------+----------+------------------------------------------------------+

启动

[root@controller ~]# nova boot --flavor m1.cirros --p_w_picpath cirros-0.3.4-x86_64 --nic net-id=a71cc567-08ad-4000-b273-e1b300fa642b \
> --security-group default --key-name demokey demo-i1

查看实例

[root@controller ~]# nova list
+--------------------------------------+---------+--------+------------+-------------+-----------------------+
| ID                                   | Name    | Status | Task State | Power State | Networks              |
+--------------------------------------+---------+--------+------------+-------------+-----------------------+
| 15a35c37-2be2-4998-b98e-e2e472df0142 | demo-i1 | ACTIVE | -          | Running     | demo-net=192.168.22.2 |
+--------------------------------------+---------+--------+------------+-------------+-----------------------+


OpenStack Icehouse私有云实战部署_linux_06

打开控制台登陆

登陆之后发现云主机并没有获取到IP，不知何故，哎呀不管了，直接手动配置

OpenStack Icehouse私有云实战部署_linux_07

进行网络连通性测试

依次ping虚拟内部网关，虚拟外部网关，真实外部网关

OpenStack Icehouse私有云实战部署_linux_08

通过以上测试发现，云主机网络正常，但是外部主机能否跟云主机通信呢？

OpenStack Icehouse私有云实战部署_美国宇航局_09

由此可以看出，外部主机还不可以与云主机通信，要想解决这一问题就需要用到floating ip机制


floating ip

简单来讲，floating ip 就是通过网络名称空间虚拟出一台路由器设备，其外部接口桥接至可通过物理接口与外部网络通信的网桥设备，而内部接口则做为内部网桥设备上关联的各虚拟机的网关接口，而后在外部网络接口上配置一个ip地址，并通过DNAT的方式转换至内部某指定的主机上，反过来，从内部某指定的主机上发出的报文则由路由器通过SNAT机制转发至外部接口上某特定的地址，从而实现了外部网络与内部VM的通信。

创建floating ip

依旧在Controller节点配置

[root@controller ~]# neutron floatingip-create ext-net
Created a new floatingip:
+---------------------+--------------------------------------+
| Field               | Value                                |
+---------------------+--------------------------------------+
| fixed_ip_address    |                                      |
| floating_ip_address | 172.16.20.13                         |
| floating_network_id | d44c19c2-2fe1-40e8-b07d-094111fe1a5e |
| id                  | de133088-d319-4094-9a2e-0b1762c85061 |
| port_id             |                                      |
| router_id           |                                      |
| status              | DOWN                                 |
| tenant_id           | 684ae003069d41d883f9cd0fcb252ae7     |
+---------------------+--------------------------------------+

将floating ip绑定至目标实例

[root@controller ~]# nova floating-ip-associate demo-i1 172.16.20.13
[root@controller ~]# nova list
+--------------------------------------+---------+--------+------------+-------------+-------------------------------------+
| ID                                   | Name    | Status | Task State | Power State | Networks                            |
+--------------------------------------+---------+--------+------------+-------------+-------------------------------------+
| 15a35c37-2be2-4998-b98e-e2e472df0142 | demo-i1 | ACTIVE | -          | Running     | demo-net=192.168.22.2, 172.16.20.13 |
+--------------------------------------+---------+--------+------------+-------------+-------------------------------------+

修改默认安全策略

[root@controller ~]# nova secgroup-add-rule default icmp -1 -1 0.0.0.0/0
+-------------+-----------+---------+-----------+--------------+
| IP Protocol | From Port | To Port | IP Range  | Source Group |
+-------------+-----------+---------+-----------+--------------+
| icmp        | -1        | -1      | 0.0.0.0/0 |              |
+-------------+-----------+---------+-----------+--------------+

现在外部网络中的主机即可通过172.16.20.13进行访问

OpenStack Icehouse私有云实战部署_虚拟主机_10

其实是由192.168.22.2进行响应的，这里就不抓包分析了

如里还需要通过ssh方式远程连接172.16.20.13，还需要执行如下命令

#nova secgroup-add-rule default tcp 22 22 0.0.0.0/0

至此，私有云基本搭建成功，接下来再说一下另一核心组件cinder，即存储服务

Block Storage服务

在没有共享存储的前提下终止实例就意味删除实例，映像文件也会被删除，要想实现用户在实例上创建的文件在实例重新创建后依然存在，只要在众compute节后背后使用共享存储即可。

Controller节点

安装所需软件包

[root@controller ~]# yum install openstack-cinder

初始化cinder数据库

[root@controller ~]# openstack-db --init --service cinder --password cinder

配置cinder服务

配置连入数据库的URL

[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf \
> database connection mysql://cinder:cinder@controller/cinder

在keystone中创建cinder用户

[root@controller ~]# keystone user-create --name cinder --pass cinder --email cinder@scholar.com
+----------+----------------------------------+
| Property |              Value               |
+----------+----------------------------------+
|  email   |        cinder@scholar.com        |
| enabled  |               True               |
|    id    | 57ec93556e744300a1f0217c26fd912b |
|   name   |              cinder              |
| username |              cinder              |
+----------+----------------------------------+
[root@controller ~]# keystone user-role-add --user=cinder --tenant=service --role=admin

连入keystone配置

[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf DEFAULT \
>   auth_strategy keystone
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_uri http://controller:5000
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_host controller
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_protocol http
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_port 35357
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_user cinder
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_tenant_name service
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_password cinder

配置其使用消息队列

[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT rpc_backend qpid
[root@controller ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT qpid_hostname controller

在keystone中注册cinder服务

[root@controller ~]# keystone service-create --name=cinder --type=volume --description="OpenStack Block Storage"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |     OpenStack Block Storage      |
|   enabled   |               True               |
|      id     | 15cbd46094f541e49f5d7a717d65101a |
|     name    |              cinder              |
|     type    |              volume              |
+-------------+----------------------------------+
[root@controller ~]# keystone endpoint-create \
>   --service-id=$(keystone service-list | awk '/ volume / {print $2}') \
>   --publicurl=http://controller:8776/v1/%\(tenant_id\)s \
>   --internalurl=http://controller:8776/v1/%\(tenant_id\)s \
>   --adminurl=http://controller:8776/v1/%\(tenant_id\)s
+-------------+-----------------------------------------+
|   Property  |                  Value                  |
+-------------+-----------------------------------------+
|   adminurl  | http://controller:8776/v1/%(tenant_id)s |
|      id     |     0e71b9f2dad24f699dce6be1ce8f40be    |
| internalurl | http://controller:8776/v1/%(tenant_id)s |
|  publicurl  | http://controller:8776/v1/%(tenant_id)s |
|    region   |                regionOne                |
|  service_id |     15cbd46094f541e49f5d7a717d65101a    |
+-------------+-----------------------------------------+
[root@controller ~]# keystone service-create --name=cinderv2 --type=volumev2 --description="OpenStack Block Storage v2"
+-------------+----------------------------------+
|   Property  |              Value               |
+-------------+----------------------------------+
| description |    OpenStack Block Storage v2    |
|   enabled   |               True               |
|      id     | dbd3b5d766f546cfb54dfc8a75f56a8e |
|     name    |             cinderv2             |
|     type    |             volumev2             |
+-------------+----------------------------------+
[root@controller ~]# keystone endpoint-create \
>   --service-id=$(keystone service-list | awk '/ volumev2 / {print $2}') \
>   --publicurl=http://controller:8776/v2/%\(tenant_id\)s \
>   --internalurl=http://controller:8776/v2/%\(tenant_id\)s \
>   --adminurl=http://controller:8776/v2/%\(tenant_id\)s
+-------------+-----------------------------------------+
|   Property  |                  Value                  |
+-------------+-----------------------------------------+
|   adminurl  | http://controller:8776/v2/%(tenant_id)s |
|      id     |     40edb783979842e99f95d75cfc5abbe8    |
| internalurl | http://controller:8776/v2/%(tenant_id)s |
|  publicurl  | http://controller:8776/v2/%(tenant_id)s |
|    region   |                regionOne                |
|  service_id |     dbd3b5d766f546cfb54dfc8a75f56a8e    |
+-------------+-----------------------------------------+

启动服务

[root@controller ~]# service openstack-cinder-api start
Starting openstack-cinder-api:                             [  OK  ]
[root@controller ~]# service openstack-cinder-scheduler start
Starting openstack-cinder-scheduler:                       [  OK  ]
[root@controller ~]# chkconfig openstack-cinder-api on
[root@controller ~]# chkconfig openstack-cinder-scheduler on

配置存储节点

准备逻辑卷

[root@block ~]# pvcreate /dev/sdb
  Physical volume "/dev/sdb" successfully created
[root@block ~]# vgcreate cinder-volumes /dev/sdb
  Volume group "cinder-volumes" successfully created

安装并配置cinder存储服务

安装所需软件包

[root@block ~]# yum install openstack-cinder scsi-target-utils

keystone相关配置

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf DEFAULT \
>   auth_strategy keystone
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_uri http://controller:5000
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_host controller
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_protocol http
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   auth_port 35357
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_user cinder
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_tenant_name service
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf keystone_authtoken \
>   admin_password cinder

消息队列配置

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT rpc_backend qpid
[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT qpid_hostname controller

连接数据库配置

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
> database connection mysql://cinder:cinder@controller/cinder

配置本节点提供cinder-volume服务使用的接口

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf DEFAULT my_ip 192.168.10.126

指定Glance服务节点

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT glance_host controller

指定卷信息文件存放位置

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
> DEFAULT volumes_dir /etc/cinder/volumes

配置scsi-target

[root@block ~]# openstack-config --set /etc/cinder/cinder.conf \
>   DEFAULT iscsi_helper tgtadm
[root@block ~]# vim /etc/tgt/targets.conf

include /etc/cinder/volumes/*

启动服务

fedora的epel源的中icehouse版本的openstack-cinder的服务openstack-cinder-volume默认为先读取/usr/share/cinder/cinder-dist.conf 这个配置文件，而其内容是有错误的。直接启动会导致创建后的卷无法关联至instace上，所以请禁止服务不再读取此文件。

[root@block ~]# service openstack-cinder-volume start
Starting openstack-cinder-volume:                          [  OK  ]
[root@block ~]# service tgtd start
Starting SCSI target daemon:                               [  OK  ]
[root@block ~]# chkconfig openstack-cinder-volume on
[root@block ~]# chkconfig tgtd on

卷创建测试

在cinder Controller节点执行如下命令，创建一个5G 大小名为demoVolume的逻辑卷

[root@controller ~]# cinder create --display-name demoVolume 5
+---------------------+--------------------------------------+
|       Property      |                Value                 |
+---------------------+--------------------------------------+
|     p_w_uploads     |                  []                  |
|  availability_zone  |                 nova                 |
|       bootable      |                false                 |
|      created_at     |      2015-07-27T15:08:11.145570      |
| display_description |                 None                 |
|     display_name    |              demoVolume              |
|      encrypted      |                False                 |
|          id         | ab0d03a8-4e89-4a17-8dc3-3432426f07a2 |
|       metadata      |                  {}                  |
|         size        |                  5                   |
|     snapshot_id     |                 None                 |
|     source_volid    |                 None                 |
|        status       |               creating               |
|     volume_type     |                 None                 |
+---------------------+--------------------------------------+

列出所有卷

[root@controller ~]# cinder list
+--------------------------------------+-----------+--------------+------+-------------+----------+-------------+
|                  ID                  |   Status  | Display Name | Size | Volume Type | Bootable | Attached to |
+--------------------------------------+-----------+--------------+------+-------------+----------+-------------+
| ab0d03a8-4e89-4a17-8dc3-3432426f07a2 | available |  demoVolume  |  5   |     None    |  false   |             |
+--------------------------------------+-----------+--------------+------+-------------+----------+-------------+

将此卷附加至指定的实例上

[root@controller ~]# nova volume-attach demo-i1 ab0d03a8-4e89-4a17-8dc3-3432426f07a2
+----------+--------------------------------------+
| Property | Value                                |
+----------+--------------------------------------+
| device   | /dev/vdb                             |
| id       | ab0d03a8-4e89-4a17-8dc3-3432426f07a2 |
| serverId | 15a35c37-2be2-4998-b98e-e2e472df0142 |
| volumeId | ab0d03a8-4e89-4a17-8dc3-3432426f07a2 |
+----------+--------------------------------------+

查看关联结果

[root@controller ~]# cinder list
+--------------------------------------+--------+--------------+------+-------------+----------+--------------------------------------+
|                  ID                  | Status | Display Name | Size | Volume Type | Bootable |             Attached to              |
+--------------------------------------+--------+--------------+------+-------------+----------+--------------------------------------+
| ab0d03a8-4e89-4a17-8dc3-3432426f07a2 | in-use |  demoVolume  |  5   |     None    |  false   | 15a35c37-2be2-4998-b98e-e2e472df0142 |
+--------------------------------------+--------+--------------+------+-------------+----------+--------------------------------------+

OpenStack Icehouse私有云实战部署_云计算_11

挂载成功，接下来就可以打开对应实例控制台，查看磁盘的附加状态，并对磁盘进行相应的操作了，这里就不再演示了

The end

喜大普奔，终于结束了，非核心组件就不做介绍了，这篇幅我也是深深的醉了，做个实验真不容易，因内存有限，死卡死卡的，真怀疑会突然卡爆掉，辛亏不是2G的RAM。以上仅为个人学习整理，如有错漏，大神勿喷~~~


转载请注明：linux运维部落  OpenStack Icehouse私有云实战部署