在OpenStack中部署kanyun

139次阅读

没有评论

共计 4075 个字符，预计需要花费 11 分钟才能阅读完成。

1 全部节点配置
- 1.1 安装 zmq
- 1.2 安装 kanyun
2 控制节点配置
- 2.1 安装配置 cassandra
- 2.2 启动和配置 kanyun
3 计算节点配置
- 3.1 启动和配置 kanyun

1 全部节点

1.1 安装 zmq

yum install libtool autoconf automake
yum install uuid-devel
yum install libuuid-devel
wget http://download.zeromq.org/zeromq-2.2.0.tar.gz
tar zxvf zeromq-2.2.0.tar.gz & cd zeromq-2.2.0 & ./configure & make install
ldconfig

easy_install“pyzmq==2.2.0”
如果出现缺少 libzmq.so.1 的情况，可能是链接库没有正确配置。
sudo /sbin/ldconfig -v grep libzmq 查看是否有所需链接
如果没有 libzmq.so.1 -> libzmq.so.1.0.1，则修改 /etc/ld.so.conf，在末尾增加 /usr/local/lib，然后 ldconfig -v grep libzmq 即可
如果出现缺少 Python.h，则使用 yum install python-devel 来安装

1.2 安装 kanyun

kanyun 原来的代码有 bug，而且不能在 python2.6 环境下使用，下面是我改过的

cd /opt
git clone https://github.com/suyan/kanyun.git

修改内容：

兼容 python2.6 和 2.7（python2.6 需要安装 OrderedDict）
修改获取内存信息方式和计算 cpu 使用率方式
修复一些 bug

python /opt/kanyun/setup.py install

mkdir /var/log/kanyun

easy_install OrderedDict

2 控制节点

2.1 安装配置 cassandra

cd /opt
wget http://mirror.bit.edu.cn/apache/cassandra/0.8.10/apache-cassandra-0.8.10-bin.tar.gz

tar zxvf apache-cassandra-0.8.10-bin.tar.gz

/opt/apache-cassandra-0.8.10/bin/cassandra

通过命令打开数据库

/opt/apache-cassandra-0.8.10/bin/cassandra-cli -h 127.0.0.1

并执行下面命令

CREATE keyspace DATA;
USE DATA;

CREATE COLUMN family vmnetwork WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family cpu WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family mem_max WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family mem_free WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family nic_incoming WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family nic_outgoing WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family blk_read WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';
CREATE COLUMN family blk_write WITH column_type='Super' AND comparator='AsciiType' AND subcomparator='IntegerType' AND default_validation_class='AsciiType';

assume vmnetwork KEYS AS ascii;
assume cpu KEYS AS ascii;
assume mem_max KEYS AS ascii;
assume nic_incoming KEYS AS ascii;
assume nic_outgoing KEYS AS ascii;
assume blk_read KEYS AS ascii;
assume blk_write KEYS AS ascii;
assume mem_free KEYS AS ascii;

easy_install pycassa

如果出现 The required version of distribute is not available

easy_install -U distribute

2.2 启动和配置 kanyun

文件中 host 需要填写计算节点的 ip
sql_connection 填写 nova.conf 中 connection 值

配置文件

[kanyun]
log: /var/log/kanyun/kanyun.log
[DEFAULT]
sql_connection: mysql://nova:novamysqlpassword@controller/nova
[server]
host: 172.19.9.1 
port: 5551
db_host: 127.0.0.1
log: /var/log/kanyun/kanyun-server.log
[api]
api_host: 172.19.9.1
api_port: 5552
db_host: 127.0.0.1
log: /var/log/kanyun/kanyun-api.log
[client]
api_host: 172.19.9.1
api_port: 5552
log: /var/log/kanyun/kanyun-client.log

python /opt/kanyun/bin/kanyun-server &
python /opt/kanyun/bin/kanyun-api &

3 计算节点

3.1 启动和配置 kanyun

文件中 api_host 需要填写计算节点的 ip
sql_connection 填写 nova.conf 中 connection 值
id 根据节点编号自己设置

配置文件

[kanyun]
log: /var/log/kanyun/kanyun.log
[DEFAULT]
sql_connection: mysql://nova:novamysqlpassword@controller/nova
[worker]
id: worker1
worker_timeout: 60
dataserver_host: 172.19.9.1
dataserver_port: 5551
log: /var/log/kanyun/kanyun-worker.log
[client]
api_host: 172.19.9.1
api_port: 5552
log: /var/log/kanyun/kanyun-client.log