Shulou(Shulou.com)05/31 Report--

What is the experience of the operation and maintenance of linux server cluster? in view of this problem, this article introduces the corresponding analysis and solution in detail, hoping to help more partners who want to solve this problem to find a more simple and feasible method.

The following is my own experience and views in the work of operation and maintenance. I hope it will be helpful to you.

1. The distinction between server models makes hardware preparations for future unification and standardization, but many people ignore this point. In fact, if this point is done well, the later operation and maintenance work will be much easier. According to the application, we mainly divide the server into three, cpu-intensive, mainly used for a large number of computing applications, such as p2p; memory-intensive, for cache applications, such as squid,varnish cache server. Disk-intensive, used for large storage applications, such as video storage servers, hadoop log storage clusters.

two。 Automatic installation of the system, mainly kickstart and cobbler

3. Unified Yum source and customized rpm package, and integrated into Yum origin server to prepare software for subsequent environment initialization

4. Build your own private network DNS

5. Standardized and unified naming (standardized basis), easy to use puppet management, and reduce operational errors, if each machine's hostname is localhost, it will be a terrible thing. The following naming methods are mainly used in our production environment

Computer room-main business-application-the last two digits of IP-company domain name, so you can see which server it is and what business it is applied to, and the alarm can also be located directly.

6. Automated configuration management and environment deployment tools: puppet,puppet module writing should minimize the direct coupling of the module, and use class inheritance to reduce the workload of operation and maintenance. Customized facter variables will make the software configuration environment more flexible. As puppet does not support clustering for the time being, multiple sets need to be deployed in practical applications. According to experience, there will be performance problems in puppet when there are about 1500 server.

7. Powerful and effective monitoring system, in the production environment, we use the cluster structure of zabbix proxy+zabbix master, zabbix can achieve effective system and application level monitoring, application monitoring also uses ppmon to achieve multi-point monitoring.

One of the biggest advantages of choosing zabbix is that the monitoring data is stored in the database, so you can use the data in the database to do a lot of operations, such as analyzing the performance indicators of the server over a period of time, checking the resource utilization of the server, and aggregating the data, so as to analyze the indicators of the whole network, such as total traffic and total http code distribution.

8. Log collection server cluster and qos analysis system, build an effective log collection system can effectively integrate and analyze users' access data, can quickly analyze qos, corresponding to important nodes we use local analysis and import mongodb, and finally import zabbix, non-important nodes directly package and compress logs and upload them to hadoop data warehouse cluster through ftp.

9. Build a redundant structure, eliminate a single point, and adopt the scheme of keepalived-ha for some important nodes in the generation environment to improve the redundancy. For resin, php and other application servers, nginx is used as a reverse proxy at the front end, while nginx uses keepalived-ha

10. Automated code distribution system, mainly the use of controltier + svn, can easily and quickly deploy code.

The answer to the question about the experience of linux server cluster operation and maintenance is shared here. I hope the above content can be of some help to you. If you still have a lot of doubts to be solved, you can follow the industry information channel for more related knowledge.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.