1. 环境准备
OS:CentOS 6.4
关闭selinux和iptables
部署Puppet:1.0 Puppet 3.7部署
安装Puppet源:http://yum.puppetlabs.com/puppetlabs-release-el-6.noarch.rpm
Dashboard是Puppet官方提供的一个简易的Web UI,可以很方便的做Puppet报告的展示和ENC的管理.但是很功能很简易。后面会更新一个更强大的Foreman,敬请期待..
2. 安装Dashboard
#安装dashboard包和mysql数据库 yum install puppet-dashboard mysql-server mysql -y /etc/init.d/mysqld start chkconfig mysqld on #初始化数据库相关... mysql -e 'CREATE DATABASE dashboard_production CHARACTER SET utf8;' mysql -e "CREATE USER 'dashboard'@'localhost' IDENTIFIED BY 'dbsa.cn';" mysql -e "GRANT ALL PRIVILEGES ON dashboard_production.* TO 'dashboard'@'localhost';" mysql -e 'set GLOBAL max_allowed_packet = 33554432;' #在/etc/my.cnf中增加以下参数 [mysqld] max_allowed_packet = 32M #执行表迁移任务,会生成相关的数据库表.. cd /usr/share/puppet-dashboard/ sed -i 's/UTC/Beijing/' config/environment.rb #修改显示时区 sed -ri 's/\s+password:.*/ password: dbsa.cn/' config/database.yml #修改默认的数据库密码.. rake RAILS_ENV=production db:migrate cp /usr/share/puppet-dashboard/ext/puppet/puppet_dashboard.rb /usr/lib/ruby/site_ruby/1.8/puppet/reports #使用默认的WEBrick服务运行dashboard测试以下... cd /usr/share/puppet-dashboard/ ./script/server -e production #访问dashboard... http://ip:3000/
3. 配置Puppet
配置由dashaboard分析Puppet的报告
#在Puppet Master上配置... /etc/puppet/puppet.conf [master] reports = store, http reporturl = http://localhost:3000/reports/upload #localhost是发往dashboard的地址
配置由dashboard管理class和parameters
#在Puppet Master上配置... /etc/puppet/puppet.conf [master] node_terminus = exec #localhost是发往dashboard的地址 external_nodes = /usr/bin/env PUPPET_DASHBOARD_URL=http://localhost:3000 /usr/share/puppet-dashboard/bin/external_node
4. 配置Nginx替代默认的WEBrick
安装Ningx和passenger
安装请参考.. 1.0 Puppet 3.7 部署中的Nginx.
Nginx配置文件:
cat > /usr/local/nginx/conf/nginx.conf <<EOF user root; worker_processes 1; events { worker_connections 1024; } http { passenger_root /usr/lib/ruby/gems/1.8/gems/passenger-3.0.19; passenger_ruby /usr/bin/ruby; passenger_max_pool_size 32; include mime.types; default_type application/octet-stream; log_format main '\$remote_addr - \$remote_user [$time_local] "\$request" ' '\$status "\$http_referer" ' '"\$http_user_agent" "\$http_x_forwarded_for"' '\$upstream_addr \$upstream_cache_status \$upstream_status'; sendfile on; keepalive_timeout 65; include /usr/local/nginx/conf/puppet.conf; #如果按照 1.0 在本机部署了Puppet这个就不用注释... include /usr/local/nginx/conf/dashboard.conf; } EOF cat > /usr/local/nginx/conf/dashboard.conf <<EOF server { listen 3000; server_name localhost; root /usr/share/puppet-dashboard/public; passenger_enabled on; access_log /usr/local/nginx/logs/access-3000.log main; } EOF #启动dashboard服务 /etc/init.d/nginx restart touch /usr/share/puppet-dashboard/log/production.log chown -R puppet-dashboard.puppet-dashboard /usr/share/puppet-dashboard/log chown -R puppet-dashboard.puppet-dashboard /usr/share/puppet-dashboard/tmp chmod 666 /usr/share/puppet-dashboard/log/production.log /etc/init.d/puppet-dashboard-workers start
报告处理过程:
1. Puppet Master在收到Agent的报告的时候,会通过http往自身的/reports/upload 接口POST一份..
2. dashboard接收到post来的报告会保存到/usr/share/puppet-dashboard/spool
3. 启动puppet-dashboard-workers 会开始处理报告..。
默认puppet-dashboard-workers 启动两个进程处理,如果报告很多, 可以再配置进程数量.
cat >> /etc/sysconfig/puppet-dashboard <<EOF CPUS=`cat /proc/cpuinfo |grep process|wc -l` EOF
4. 处理完成的报告会写入数据库
5. 节点管理(ENC)
在前面的master puppet.conf中配置使用enc。
#下面是一个测试.. cat > /etc/puppet/manifests/site.pp <<EOF node default { } EOF
1. 在Dashboard 添加一个class。。(base的class在之前的blog中已经创建过,可以查看1.0 Puppet安装)
2. 选择一个主机..
3. 给这个主机添加class和parameters
4. 查看添加的类和参数..
5. 通过执行命令,查询配置
[root@master ~]# /usr/bin/env PUPPET_DASHBOARD_URL=http://localhost:3000 /usr/share/puppet-dashboard/bin/external_node agent1.dbsa.cn --- classes: - base parameters: aa: bb name: agent1.dbsa.cn
6. 常用维护命令
导入现有的日志
cd /usr/share/puppet-dashboard;rake RAILS_ENV=production reports:import cd /usr/share/puppet-dashboard;rake RAILS_ENV=production reports:import REPORT_DIR=/path/to/your/reports
当数据量过大时,优化数据库
cd /usr/share/puppet-dashboard;rake RAILS_ENV=production db:raw:optimize
删除一个月之前的日志
cd /usr/share/puppet-dashboard;rake RAILS_ENV=production reports:prune upto=1 unit=mon
删除15天前的日志
cd /usr/share/puppet-dashboard;rake RAILS_ENV=production reports:prune upto=15 unit=day
备份数据库
mysqldump --add-locks --create-options --disable-keys --extended-insert --quick --set-charset --user=dashboard --password=dbsa.cn dashboard > production.sql
恢复数据库
rake RAILS_ENV=production FILE=production.sql db:raw:restore