About centos-admin.ru
Today, the system administration service of servers is one of the most popular, because each of the websites operating on the Internet works on a server or servers that one way or another someone administers.
Different system administrators like different operating systems - to whom that is more familiar and more convenient: CentOS Linux (or Red Hat Linux), Fedora Core, Debian, Ubuntu, FreeBSD, Microsoft Windows Server and many others. You can reason a lot, which of the systems is faster and more stable, but, in my opinion, what’s important is not the choice of the operating system, but the experience of the system administrator, the ability to work in a team.
My choice fell on CentOS + OpenVZ virtualization. CentOS - because it is convenient to administer it on a large number of servers (our team currently supports more than 300 physical servers, each of which runs several virtual containers). OpenVZ - because it is very convenient and reliable virtualization with a minimum overhead of resources.
We squeeze the maximum out of CentOS. For maximum performance, we fine-tune the network subsystem, where necessary - update the drivers, tighten the sysctl-parameters.
And virtualization gives us the possibility of hot backups (including remote ones in case the server burns down), fast transfer of software and data between hardware (including minimal or even zero downtime), fast scaling of loaded projects (you can compare it with a modular word “Cloud” :-).
')
We set up servers so that websites open as quickly as possible, and the possible attendance was as large as possible. Properly configured software such as nginx, apache, mysql, postgresql, php, ruby, use of accelerators like php-pecl-apc, proper work with statics, clever limit on the number of processors and many other subtleties we take into account help to achieve high-quality and stable operation of servers.
Monitoring plays an important role in our work. We try to keep track of all sorts of parameters: temperature of the processor, hard drives, control of loading data transfer channels, responsiveness of the site and much more. Upon request, we provide access to schedules of CPU usage, memory, the number of requests to the web server, database, etc ... And most importantly! We try to predict the problem and solve it in advance - we love to sleep at night :-) It's easy to see that the hard drive is about to burn out and pre-order its replacement at the time agreed with the client and the DC (hoster).
If you need professional server maintenance, we will be glad to see you among our clients!
Different system administrators like different operating systems - to whom that is more familiar and more convenient: CentOS Linux (or Red Hat Linux), Fedora Core, Debian, Ubuntu, FreeBSD, Microsoft Windows Server and many others. You can reason a lot, which of the systems is faster and more stable, but, in my opinion, what’s important is not the choice of the operating system, but the experience of the system administrator, the ability to work in a team.
My choice fell on CentOS + OpenVZ virtualization. CentOS - because it is convenient to administer it on a large number of servers (our team currently supports more than 300 physical servers, each of which runs several virtual containers). OpenVZ - because it is very convenient and reliable virtualization with a minimum overhead of resources.
We squeeze the maximum out of CentOS. For maximum performance, we fine-tune the network subsystem, where necessary - update the drivers, tighten the sysctl-parameters.
And virtualization gives us the possibility of hot backups (including remote ones in case the server burns down), fast transfer of software and data between hardware (including minimal or even zero downtime), fast scaling of loaded projects (you can compare it with a modular word “Cloud” :-).
')
We set up servers so that websites open as quickly as possible, and the possible attendance was as large as possible. Properly configured software such as nginx, apache, mysql, postgresql, php, ruby, use of accelerators like php-pecl-apc, proper work with statics, clever limit on the number of processors and many other subtleties we take into account help to achieve high-quality and stable operation of servers.
Monitoring plays an important role in our work. We try to keep track of all sorts of parameters: temperature of the processor, hard drives, control of loading data transfer channels, responsiveness of the site and much more. Upon request, we provide access to schedules of CPU usage, memory, the number of requests to the web server, database, etc ... And most importantly! We try to predict the problem and solve it in advance - we love to sleep at night :-) It's easy to see that the hard drive is about to burn out and pre-order its replacement at the time agreed with the client and the DC (hoster).
If you need professional server maintenance, we will be glad to see you among our clients!
Source: https://habr.com/ru/post/159275/
All Articles