puppet在比较老的版本的时候是通过kick进行推送实现配置及时更新,由于kick的效率问题,在比较新的版本中开始采用第三方工具MCollective来实现,网上介绍如何部署MCollective文章也不少,但大都是linux平台下的部署,windows下的几乎没有,我在研究的时候也是一路坎坷,差点放弃,不过看到有一个老外自己成功部署后,有点不甘心,在坚持下终于配置成功,现在整理一下分享出来。
如果你还一点都不了解MCollective,建议看看这篇文章,他讲的是在linux下部署MCollective,其中讲了各个工具的关系,其实流程都一样,只不过我这里讲的是windows客户端的部署。
=========================================================================================
以下每个用到的工具我都有写明自己部署时候的版本号,如果你部署完成后无法正常实现推送,可以重新指定安装每个工具的版本
1、首先配置好puppet server和puppet client,这个很容易实现,就不详细说明了,我用到版本分别是
puppet server(centos 6.2): 3.4.3
puppet client(windows 2008 r2): 3.4.3
2、安装一款 消息队列服务,你可以选择RabbitMQ或者ActiveMQ,我这里选择的是ActiveMQ,ActiveMQ是装在puppet server端,使用如下命令进行安装,
当前ActiveMQ版本号为:5.8.0
yum install tanukiwrapper activemq activemq-info-provider
安装完成后来对ActiveMQ进行配置,如下
[root@puppetserver rpms]# vim /etc/activemq/activemq.xml … <simpleAuthenticationPlugin> <users> <!-- <authenticationUser username="${activemq.username}" password="${activemq.password}" groups="admins,everyone"/> --> #禁用 <authenticationUser username="mcollective" password="secret" groups="mcollective,admins,everyone"/> #配置通信的账号及密码 </users> </simpleAuthenticationPlugin> … <authorizationPlugin> #配置权限,默认即可 <map> <authorizationMap> <authorizationEntries> <authorizationEntry queue=">" write="admins" read="admins" admin="admins" /> <authorizationEntry topic=">" write="admins" read="admins" admin="admins" /> <authorizationEntry topic="mcollective.>" write="mcollective" read="mcollective" admin="mcollective" /> <authorizationEntry topic="mcollective.>" write="mcollective" read="mcollective" admin="mcollective" /> <authorizationEntry topic="ActiveMQ.Advisory.>" read="everyone" write="everyone" admin="everyone"/> </authorizationEntries> </authorizationMap> </map> </authorizationPlugin> … <transportConnectors> <transportConnector name="openwire" uri="tcp://0.0.0.0:61616"/> <transportConnector name="stomp+nio" uri="stomp://0.0.0.0:61613"/> #配置通信协议为stomp,监听61613端口 </transportConnectors>
然后启动ActiveMQ
[root@puppetserver rpms]# service activemq start Starting ActiveMQ Broker... [root@puppetserver rpms]# chkconfig activemq on [root@puppetserver rpms]# netstat -nlatp | grep 61613 #查看监听端口 tcp 0 0 :::61613 :::* LISTEN 33805/java
3、现在开始配置MCollective,MCollective也分为客户端和服务器端,其客户端是装在puppetserver上,服务器端是装下puppetclient,我们先配置客户端,也就是puppetserver上的环境,命令如下,当前mcollective客户端版本是2.5.0:
[root@puppetserver rpms]# yum install mcollective-common mcollective-client
然后配置mcollective客户端:
[root@puppetserver rpms]# vim /etc/mcollective/client.cfg main_collective = mcollective collectives = mcollective libdir = /usr/libexec/mcollective logger_type = console loglevel = warn # Plugins securityprovider = psk plugin.psk = test123123 #MCollective通信共享密钥,和MCollective服务端保持一致 connector = activemq #MCollective用到的通讯组件 plugin.activemq.pool.size = 1 plugin.activemq.pool.1.host = 192.168.3.91 #本机ip plugin.activemq.pool.1.port = 61613 #AcitveMQ中设置的端口号 plugin.activemq.pool.1.user = mcollective #AcitveMQ中设置的用户 plugin.activemq.pool.1.password = secret #AcitveMQ中设置的密码 plugin.activemq.heartbeat_interval = 30 # Facts factsource = yaml plugin.yaml = /etc/mcollective/facts.yaml
接下来是个需要注意的地方,mcollective本身并没有相应的windows服务端安装包,在windows下安装mcollective是一件麻烦的事情,不过有大神做了mcollective安装程序,它集成了mcollective运行所需要的一个gem包,以及将mcollective安装成windows服务,点击这里下载安装包,我下载的是2.3.2版本,下载后在puppet client端安装,安装完成后windows服务里会有一个服务名:The Marionette Collective。
注意安装前检查如下环境变量,安装的时候才用右键管理员运行程序包:
Path变量是否包含 D:Puppet LabsPuppetsys ubyin;D:Puppet LabsPuppetin
然后配置mcollective服务端server.cfg,我的目录是D:mcollectiveetc下:
main_collective = mcollective collectives = mcollective libdir = d:mcollectiveplugins #mcollective插件目录 logfile = d:mcollectivemcollective.log #mcollective日志 loglevel = info daemonize = 1 # Plugins securityprovider = psk plugin.psk = test123123 connector = activemq plugin.activemq.pool.size = 1 plugin.activemq.pool.1.host = 192.168.3.91 plugin.activemq.pool.1.port = 61613 plugin.activemq.pool.1.user = mcollective plugin.activemq.pool.1.password = secret plugin.activemq.pool.1.heartbeat_interval = 30 # Facts factsource = yaml plugin.yaml = d:mcollectiveetcfacts.yaml
完成后在puppetclient上启动mcollective服务,此时mcollective的服务端和客户端配置就已经成功,可以在puppetserver上测试mcollective的通讯是否正常,如下
[root@puppetserver mcollective]# mco ping waiwofei time=42.00 ms WIN-6BJIEV6N3T6 time=80.98 ms WIN-3JOT1GVGEOO time=89.42 ms ---- ping statistics ---- 3 replies max: 89.42 min: 42.00 avg: 70.80
4、为mcollective安装puppet agent插件,首先在puppet server端安装,当前版本是1.7.1
yum install mcollective-puppet-client mcollective-puppet-common
接着要在puppet client安装,如果是linux可以通过命令向上面那样直接安装,window的话我们只能下载插件包,然后把相关文件复制到mcollective插件目录,通过下面地址下载,我最初下载的是1.7.1版本,但是在mcollective启动的时候无法正常加载puppet agent插件,后来换成1.6.1就没问题了,版本的问题困扰了我很久,几乎要放弃了,但是看到某个老外配置成功,发现帖子的时间是2013年6月份,我就索性把插件版本降到比较老,结果正常了。好了,通过下面下载
https://github.com/puppetlabs/mcollective-puppet-agent/releases
解压后将里面的文件夹复制到mcollective的插件目录,我的插件目录[d:mcollectiveplugins],接着再次配置puppet client端mcollective的server.cfg,增加如下两行,这个插件的配置写法也卡住了我很久,说多了都是泪啊。。。。
main_collective = mcollective collectives = mcollective libdir = d:mcollectiveplugins logfile = d:mcollectivemcollective.log loglevel = info daemonize = 1 # Plugins securityprovider = psk plugin.psk = test123123 connector = activemq plugin.activemq.pool.size = 1 plugin.activemq.pool.1.host = 192.168.3.91 plugin.activemq.pool.1.port = 61613 plugin.activemq.pool.1.user = mcollective plugin.activemq.pool.1.password = secret plugin.activemq.pool.1.heartbeat_interval = 30 plugin.puppet.command = "D:Puppet LabsPuppetinpuppet.bat" agent plugin.puppet.config = C:ProgramDataPuppetLabspuppetetcpuppet.conf # Facts factsource = yaml plugin.yaml = d:mcollectiveetcfacts.yaml
ok,重启puppet client端mcollective服务,可以在puppet server通过下面的命令查看puppet client的插件的加载情况。
[root@puppetserver mcollective]# mco inventory waiwofei Inventory for waiwofei: Server Statistics: Version: 2.3.2 Start Time: Mon Apr 28 17:24:51 +0800 2014 Config File: D:mcollectiveetcserver.cfg Collectives: mcollective Main Collective: mcollective Process ID: 73044 Total Messages: 4 Messages Passed Filters: 4 Messages Filtered: 0 Expired Messages: 0 Replies Sent: 3 Total Processor Time: 1.014 seconds System Time: 0.39 seconds Agents: discovery puppet rpcutil Data Plugins: agent fstat puppet resource Configuration Management Classes: No classes applied Facts: mcollective => 1
注意,如果你重启不了或者puppet插件正常加载(也可以通过mcollective.log查看加载是否有错误),那么你需要检查下你的环境变量
RUBYLIB是否存在以及包含 D:Puppet LabsPuppetpuppetlib;D:Puppet LabsPuppetfacterlib
整个配置过程结束,在puppetserver上进行一次推送试试吧
[root@puppetserver mcollective]# mco puppet -v -I waiwofei runonce * [ ============================================================> ] 1 / 1 waiwofei : OK {:summary=> "Started a background Puppet run using the '"D:\Puppet Labs\Puppet\bin\puppet.bat" agent --onetime --color=false --splay --splaylimit 30' command"} ---- rpc stats ---- Nodes: 1 / 1 Pass / Fail: 1 / 0 Start Time: Tue Apr 29 10:36:58 +0800 2014 Discovery Time: 0.00ms Agent Time: 50.93ms Total Time: 50.93ms
运行成功,通过puppet dashboard也可以看到运行成功