Http://blog.sina.com.cn/s/blog_4e424e210100pboy.html
As businesses become more reliant on it, the scale of the enterprise's IT infrastructure is expanding. As a system engineer, how will we respond to this growing IT architecture? In the past, the system administrator was like a worker on the assembly line, repeating the same job; now that's all about to change, we're going to introduce a bunch of tools that are "programmable", and you just need to write a few lines of "code" for this batch of tools, which will help you do all the work automatically. And these tools are the objects of our research, and then I'll cover them in a series of articles. I write slowly, you look slowly.
the automated operations tools available to system engineers can be categorized into three main categories: preparation, configuration management, and monitoring.
In the early days, Linux administrators might have compiled a list of packages with RPM for bulk software installation. Later, we used kickstart to perform an unattended Linux installation. Now, cobbler this feature to a new level: It enables parallel system construction of physical and virtual machines, and can be configured for DHCP and DNS. Cobbler also integrates other tools, such as software for configuration Management Automation puppet. This tool can automatically update the service after the system is installed.
The Configuration management tool is used to set parameters or to open a service on a new server. Configuration management can be used for automatic server building. The server auto-build tool accelerates deployment and can deploy servers on a large scale in a short period of time, as well as making the build process easier to replicate. In the event of a serious failure, you can also rebuild the schema.
Most of these tools are real-time and can be used to make changes and perform certain tasks. However, they lack information about the current system situation, and the monitoring tools are available. For a traditional system administrator, monitoring is nothing more than a page or an e-mail alert when an error occurs. However, monitoring tools-such as Nagios, OpenNMS, and Zenoss core---can provide better server performance testing. This information tells administrators what is going on in the running application. Some of them can even invoke the configuration management tool for self-healing. For example, Zenoss Core can reconfigure services with Cfengine, chef, or puppet on a single detection status table. A scenario for automatic fault detection and failure repair is demonstrated on the DevOps gameday. A Web application in the Amazon EC2 Data Center, which is hosted both on the East and west coasts. Administrators at the West Coast have made some server failures, Zenoss Core has monitored the architecture, found faults and informed Opscode Chef to take action. Chef then updates the services on the Dynect platform via the Dynect API, which automatically restores the East Coast application.
As demonstrated on the DevOps gameday, if the entire IT architecture is self-sustaining, system engineers will no longer have to climb up at night to deal with failures, which would be a wonderful thing.
Operation and Maintenance automation
