10月15日，OpenStack发布了Liberty版本，其中一个变化就是容器即服务模块Magnum正式推出1.0版本，只是各大Linux发行厂商还没有发布magnum的安装包，因此需要从源码直接构建。下面给出自己基于Devstack方式的完整安装实践过程。

环境准备

在VMware Workstations建台虚拟机，Ubuntu 14.04 LTS，CPU两个核以上，内存至少4G以上，硬盘100G，两个网卡（一个主机模式，另一个NAT模式）。

$ vi /etc/network/interfaces
#仅主机模式，用于openstack内部模块通信
auto eth0
iface eth0 inet staticaddress 20.0.0.11gateway 20.0.0.1netmask 255.255.255.0#NAT模式，用于虚拟机访问外网的外部网络
auto eth1
iface eth1 inet staticaddress 172.24.54.222gateway 172.24.54.2netmask 255.255.255.0#将缺省路由设置为网卡2
up route add default gw 172.24.54.2 dev eth1

部署过程

1.安装git

$ sudo su
# apt-get install git

2.下载devstack源码

# cd /home
# git clone https://github.com/openstack-dev/devstack

3.创建stack用户运行

# cd /home/devstack/tools/
# ./create-stack-user.sh

4.修改devstack目录权限,让stack用户可以运行

# chown -R stack:stack /home/devstack

5.切换到stack用户，进入devstack目录下，创建localrc文件

# su stack
$ cd /home/devstack
$ cat > localrc << END
DATABASE_PASSWORD=password
RABBIT_PASSWORD=password
SERVICE_TOKEN=password
SERVICE_PASSWORD=password
ADMIN_PASSWORD=password
PUBLIC_INTERFACE=eth1
enable_plugin magnum https://git.openstack.org/openstack/magnum
enable_plugin barbican https://git.openstack.org/openstack/barbican
VOLUME_BACKING_FILE_SIZE=20G
HOST_IP=20.0.0.11
END

6.运行Devstack

$ ./stack.sh

注意：使用的是stack用户运行。在安装过程中，可能会提示apt-get下载源错误，重复执行上述安装命令；其他情况，可以再次执行安装命令。

$ ./unstack.sh && ./stack.sh

7.验证

整个安装过程耗时一个小时左右，取决于你的网络状况。在浏览器中打开http://20.0.0.11/ 即可访问horizon。默认Devstack创建admin和demo两个用户，密码是第5步中localrc中设置的password。

或者可以使用OpenStack的命令行工具来验证

$ cd /home/devstack
$ source openrc admin admin
$ glance image-list
+--------------------------------------+---------------------------------+
| ID                                   | Name                            |
+--------------------------------------+---------------------------------+
| 148f3031-13cd-4a93-85fe-03b42b2117c6 | cirros-0.3.4-x86_64-uec         |
| c291cb0c-6265-4be9-8eb2-d2988bda203c | cirros-0.3.4-x86_64-uec-kernel  |
| 502aca3d-4e41-4e41-ad50-baa82c0a6cde | cirros-0.3.4-x86_64-uec-ramdisk |
| b89a7a9f-3b2f-4291-8327-60c735495b09 | fedora-21-atomic-5              |
+--------------------------------------+---------------------------------+

窥探Magnum

Magnum一般有两个子模块，magnum-api和magnum-conductor,为了验证conductor服务是否健康运行

$ magnum service-list
+----+-----------+------------------+-------+
| id | host      | binary           | state |
+----+-----------+------------------+-------+
| 1  | localhost | magnum-conductor | up    |
+----+-----------+------------------+-------+

1.创建密钥对

$ test -f ~/.ssh/id_rsa.pub || ssh-keygen -t rsa -N "" -f ~/.ssh/id_rsa
$ nova keypair-add --pub-key ~/.ssh/id_rsa.pub testkey

2.创建BayModel对象

$ magnum baymodel-create --name k8sbaymodel \--image-id fedora-21-atomic-5 \--keypair-id testkey \--external-network-id public \--dns-nameserver 8.8.8.8 \--flavor-id m1.small \--docker-volume-size 5 \--network-driver flannel \--coe kubernetes

3.创建Bay对象

$ magnum bay-create --name k8sbay --baymodel k8sbaymodel --node-count 1

Bays对象初始化为CREATE_IN_PROGRESS状态，当创建完成后会更新为CREATE_COMPLETE状态. 此步骤耗时10多分钟，可以通过以下命令查看已有的Bays对象

$ magnum bay-list
+--------------------------------------+--------+------------+--------------+-----------------+
| uuid                                 | name   | node_count | master_count | status          |
+--------------------------------------+--------+------------+--------------+-----------------+
| f0ce1f2d-a1a9-4a47-977f-15244625463d | k8sbay | 1          | 1            | CREATE_COMPLETE |
+--------------------------------------+--------+------------+--------------+-----------------+

上述的Bay对象将会创建一个Kubernetes master节点和一个minion节点:

$ nova list
+--------------------------------------+-------------------------------------------------------+--------+------------+-------------+------------------------------+
| ID                                   | Name                                                  | Status | Task State | Power State | Networks                     |
+--------------------------------------+-------------------------------------------------------+--------+------------+-------------+------------------------------+
| 4656ff13-9210-47b5-9c27-8c26309d5e04 | k8-fpoxo3qagq-0-bax3fvtb5pf2-kube_master-riv2fk732dbd | ACTIVE | -          | Running     | private=10.0.0.5, 172.24.4.5 |
| 45b7c2af-29de-4f59-b5ba-b784e0ef503b | k8-hdx2ewllqx-0-5vptn5l4ahtt-kube-minion-i6jyfvsvxkzu | ACTIVE | -          | Running     | private=10.0.0.6, 172.24.4.6 |
+--------------------------------------+-------------------------------------------------------+--------+------------+-------------+------------------------------+

更直观的，可以在Horizon的编配里面查看相关栈或Bay的详情 

4.下载Kubernetes源码

$ wget https://github.com/kubernetes/kubernetes/releases/download/v1.0.1/kubernetes.tar.gz
$ tar -xvzf kubernetes.tar.gz

5.部署Redis容器集群

$ cd kubernetes/examples/redis
$ magnum pod-create --manifest ./redis-master.yaml --bay k8sbay$ magnum coe-service-create --manifest ./redis-sentinel-service.yaml --bay k8sbay$ sed -i 's/\(replicas: \)1/\1 2/' redis-controller.yaml
$ magnum rc-create --manifest ./redis-controller.yaml --bay k8sbay$ sed -i 's/\(replicas: \)1/\1 2/' redis-sentinel-controller.yaml
$ magnum rc-create --manifest ./redis-sentinel-controller.yaml --bay k8sbay

这样我们就创建了一个redis的ReplicationController，由这个Controller来调度和管理redis容器，通过magnum命令可以查看IP与状态。

$ magnum bay-show k8sbay
+--------------------+------------------------------------------------------------+
| Property           | Value                                                      |
+--------------------+------------------------------------------------------------+
| status             | CREATE_COMPLETE                                            |
| uuid               | f0ce1f2d-a1a9-4a47-977f-15244625463d                       |
| status_reason      | Stack CREATE completed successfully                        |
| created_at         | 2015-10-20T12:56:07+00:00                                  |
| updated_at         | 2015-10-20T13:13:41+00:00                                  |
| bay_create_timeout | 0                                                          |
| api_address        | https://172.24.4.4:6443                                    |
| baymodel_id        | 33d8e94d-f867-437f-80b4-10d5d6885c8a                       |
| node_count         | 1                                                          |
| node_addresses     | [u'172.24.4.6']                                            |
| master_count       | 1                                                          |
| discovery_url      | https://discovery.etcd.io/6533dd3e28c0ff7e410cff633fad90ad |
| name               | k8sbay                                                     |
+--------------------+------------------------------------------------------------+

注意：使用Kubernetes部署Pod时会从Google container registry下载基础的Pause镜像，但众所周知由于国内特殊的网络环境，会导致容器无法部署成功。所以咱最好还是在主机上拨个VPN。

6.验证

剩下来就是ssh到相应的虚拟机中，通过docker命令或redis客户端来控制和访问容器了。

$ ssh minion@172.24.4.6
[minion@k8-hdx2ewllqx-0-5vptn5l4ahtt-kube-minion-i6jyfvsvxkzu ~]$ sudo su
bash-4.3# docker ps
CONTAINER ID        IMAGE                                  COMMAND             CREATED             STATUS              PORTS               NAMES
6522a94e8c2a        gcr.io/google_containers/pause:0.8.0   "/pause"            58 minutes ago      Up 55 minutes                           k8s_POD.49eee8c2_redis-bldhr_default_8f5a1e81-7739-11e5-8d21-fa163ec6c383_cb8d1102
277bbcfca7df        gcr.io/google_containers/pause:0.8.0   "/pause"            58 minutes ago      Up 56 minutes                           k8s_POD.39750b55_redis-master_default_690a053d-7739-11e5-8d21-fa163ec6c383_1da0a23d
fbc0bf70079e        gcr.io/google_containers/pause:0.8.0   "/pause"            58 minutes ago      Up 56 minutes                           k8s_POD.ecf0e8f4_redis-sentinel-ii2go_default_9ca36eec-7739-11e5-8d21-fa163ec6c383_9a95d2d0