All computer secrets for beginners and professionals / Lifehacks / Automatic OS installation on dedicated servers. Installing the OS on the server (installimage) Installing the operating system using an autoinstall package

Automatic OS installation on dedicated servers. Installing the OS on the server (installimage) Installing the operating system using an autoinstall package

15.12.2023

Installing an OS on a server is one of the most critical stages when starting it. The choice of OS largely determines its performance and successful functioning.

What requirements must a server OS meet?

The server operating system must meet a number of requirements, namely:

Work stability. The operating system must be able to maintain functionality for a long time without freezing or slowing down.
Reliability. The operating system itself or with the help of antiviruses must provide reliable protection against viruses and hacker attacks.
High performance. The operation of the operating system should not take up much server capacity. The more the OS loads the server, the less capacity is left for its main work.

To one degree or another, every system intended for installation on a server meets these requirements.

What are the types of server OS?

Today, there are two types of such operating systems that are most popular on the market: Microsoft OS and Unix-based OS. An approximate list looks like this:

Windows Server. Commercial OS from Microsoft. It has high performance and reliability, but consumes a lot of hardware resources.
Red Hat Enterprise Linux. Another commercial Unix-based system. Has high reliability indicators. As with all Unix systems, it is true for Red Hat that installing an operating system on a server is a rather complex and painstaking process.
CentOS. Free analogue of the previous option. The main disadvantage is that installation and configuration occurs at your own peril and risk. Since CentOS does not have an official support service.
Debian, Ubuntu. The most popular free server OS. They are distinguished by their reliability and the fact that installation and configuration for them is a relatively simple process.
FreeBSD is an “old school” OS. Used on many services with a multi-million audience. Now this OS is not in great demand.

In addition to the presented advantages and disadvantages, in order to select and install an OS, you need to take into account many more parameters, depending on your goals.

Who installs the OS on the server?

Installing an operating system on a server is very different from installing an OS on a regular PC. To do this, you can involve the following specialists:

Full-time system administrator.
Private OS configuration specialist.
A company that specializes in installing OS on a server.

The operation of the entire server, the security of your information and the prestige of the company depend on the correct installation and configuration of your server, therefore, only professionals should be involved in setting up your server.

Using the YouDo online exchange, you can order the services of specialists who will help you install the OS efficiently, on time and at the best price. To do this, you just need to leave your phone number on this website, our employees will call you back and help you select specialists of the highest level.

The most radical way to get rid of “garbage” in the operating system is to “clean” reinstall it. Which is a fairly simple task if you have physical access to the server (computer). However, most WEB servers are located many, many kilometers away from their owners. In this article I will tell you in detail how to install CentOS remotely using only access to the server console via SSH.

The need to reinstall the operating system

There may be several reasons. Problems with operation, when the system slows down, or suddenly lose performance. From time to time some software may glitch or refuse to work at all. If you need to change the bit depth, for example from 32 bits to 64 or vice versa. When a clean OS is needed, for example, to install the ISPmanager or Vesta Panel hosting control panel.

Three ways to install an operating system on a server

Come to the DATA center and install the OS;
Pay for administration services and DATA center employees will install the operating system for you;
Use VNC and install OS remotely.

I like the third option.

Why the hell should a goat button accordion? Or why it’s better to install the OS yourself.

I see several reasons for myself to reinstall the operating system myself:

High cost of administration services. If you have a budget server and the administration service is paid separately.
Lack of control over administrator actions. I don’t trust anyone with the server or my wife 🙂
A waste of time, the hosting provider administrator may have a lot to do with other clients and their servers.
Operating system images of unknown origin. As a rule, no one will install an operating system for you from scratch from a distribution kit, but will upload a pre-prepared image to disk. Which may be far from ideal.

Three components of success

Access to the server via SSH with ROOT rights.
“Live” CentOS server with Internet access.
Software for remote access to the computer desktop, for example TightVNC

INSTRUCTIONS for remote installation of CentOS

Foreplay or where to start

I wrote, I write and I will write: don’t forget about backup! And although this is not what the article is about, it is imperative to make backup copies, since after reinstalling the operating system, your server will be as clean as a white sheet.

You need to determine the network settings of your server. We are interested in the following values:
Network interface (MAC address or name);
IP address of the server network interface;
Network mask;
Default gateway IP address;
IP address of an available DNS server, as a rule, you can use public DNS from Google 8.8.8.8 and 8.8.4.4
To determine the necessary parameters, run several commands in the console:

Ifconfig ip route show cat /etc/sysconfig/network-scripts/ifcfg-eth0 cat /etc/sysconfig/network cat /etc/resolv.conf

I got the following values:

//Ip 193.170.128.128 //Gateway 193.170.128.1 //DNS 193.170.128.2 //MASK 255.255.252.0 //MAC 12:14:01:4a:25:b5

Loading images to get started

You must select downloadable images depending on the version of the operating system and its bit depth. For example, for CentOS 64 bit version 6.4, run the following commands:

Wget -O /boot/vmlinuz_remote http://mirror.centos.org/centos/6.4/os/x86_64/isolinux/vmlinuz wget -O /boot/initrd_remote.img http://mirror.centos.org/centos/6.4 /os/x86_64/isolinux/initrd.img

If you need to install a 32-bit operating system, then change the addresses x86_64 on i386:

Wget -O /boot/vmlinuz_remote http://mirror.centos.org/centos/6.4/os/i386/isolinux/vmlinuz wget -O /boot/initrd_remote.img http://mirror.centos.org/centos/6.4 /os/i386/isolinux/initrd.img

Please note that the download will be made from official CentOS servers. First check the correctness of the paths by opening the addresses in the browser.

GRUB bootloader configuration for remote installation of CentOS

Let's move on to the most important point - setting up the GRUB bootloader configuration. In a nutshell, we will specify in the settings an alternative download using a freshly downloaded distribution. And tell grub to try loading it once. And if something goes wrong, we will automatically return to the previously installed distribution after a reboot after 120 seconds (in some cases, the reboot may require the help of DATA Center specialists).
Open the grub.conf file (usually /boot/grub/grub.conf) and add the following lines to it:

Title Remote Install root (hd0,0) kernel /boot/vmlinuz_remote lang=en_US keymap=us method=http://mirror.centos.org/centos/6.4/os/x86_64/ vnc vncpassword=123456 ip=193.170.128.128 netmask =255.255.252.0 gateway=193.170.128.1 dns=193.170.128.2 noselinux ksdevice=eth0 headless xfs panic=120 initrd /boot/initrd_remote.img

Where, root(hd0,0)- location /BOOT partition, vncpassword— password for accessing the remote desktop of the VNC server (create your own), ip— IP address of your server, netmask- network mask, gateway— IP address of the default gateway, dns— IP address of the DNS server (you can use public ones from Google 8.8.8.8 or 8.8.4.4), ksdevice— name of the network interface or its MAC address, panic— reboot time if something goes wrong.
In addition, check the download addresses for availability and the names of the files that we downloaded in the previous step.
For a 32-bit operating system, the changes will look like this (as we always change the address x86_64 to i386):

Title Remote Install root (hd0,0) kernel /boot/vmlinuz_remote lang=en_US keymap=us method=http://mirror.centos.org/centos/6.4/os/i386/ vnc vncpassword=123456 ip=193.170.128.128 netmask =255.255.252.0 gateway=193.170.128.1 dns=193.170.128.2 noselinux ksdevice=eth0 headless xfs panic=120 initrd /boot/initrd_remote.img

Now the most important thing is where to insert all this :)
My grub.conf file looks like this:

# grub.conf generated by anaconda # # Note that you do not have to rerun grub after making changes to this file # NOTICE: You have a /boot partition. This means that # all kernel and initrd paths are relative to /boot/, eg. # root (hd0,0) # kernel /vmlinuz-version ro root=/dev/mapper/VolGroup-lv_root # initrd /initrd-version.img #boot=/dev/sda default=0 timeout=5 splashimage=(hd0, 0)/grub/splash.xpm.gz hiddenmenu title CentOS (2.6.32-358.2.1.el6.i686) root (hd0,0) kernel /boot/vmlinuz-2.6.32-358.2.1.el6.i686 ro root=/dev/mapper/VolGroup-lv_root rd_NO_LUKS LANG=en_US.UTF-8 rd_NO_MD rd_LVM_LV=VolGroup/lv_swap SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup/lv_root KEYBOARDTYPE=pc KEYTABLE=us rd_NO_DM rhgb quiet initrd /boot/ initramfs-2.6.32-358.2.1.el6.i686.img title CentOS (2.6.32-358.el6.i686) root (hd0,0) kernel /boot/vmlinuz-2.6.32-358.el6.i686 ro root=/dev/mapper/VolGroup-lv_root rd_NO_LUKS LANG=en_US.UTF-8 rd_NO_MD rd_LVM_LV=VolGroup/lv_swap SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup/lv_root KEYBOARDTYPE=pc KEYTABLE=us rd_NO_DM rhgb quiet initrd /boot/ initramfs-2.6.32-358.el6.i686.img

As you can see, the system has two options in the download list. And we have to add our own. Let's add it at the very end:

Our CentOS boot configuration comes in third. Let's tell grub to try loading it once:

Grub grub> savedefault --default=2 --once savedefault --default=2 --once grub> quit

Then you can reboot the server:

We wait for the server to start pinging and try to connect via VNC at 193.170.128.128:1 (address of our server, VNC port=1). Depending on the server power and channel speed, this may take up to half an hour. During this time, all necessary packages will be downloaded and installed.
After this, the screen will display the standard CentOS installer desktop with a graphical interface. Well, installing the operating system on the server itself is a completely different story.

Notes on remote installation of CentOS

VNC does not support Keepalives and may fail if left idle for a long time. Therefore, if you have connected a VNC client, then you need to get down to business right away, or forcefully disconnect from the server.
If you do not have a grub.conf file on your server, this means that you have a virtual server based on OpenVZ and these instructions do not suit you. This is another reason to use hardware virtualization methods, for example VmWare, as done by Adman.
The article provides a safe boot method, with the ability to revert to an existing operating system. But you can write the new configuration in first place and then there will be no need for the command savedefault --default=2 --once. Grub will use the first one in the list.

Sources for inspiration when writing this article

As is already clear from the name itself, it is the main shell for the connecting link between the central software or hardware-software part of a local or global network and child client terminals. The understanding of this term has a broader interpretation, so it is worth dwelling on some theoretical and practical issues. At the very least, you need to understand the purpose of the server operating system, as well as some aspects of its installation and practical use. Users who believe that this issue relates exclusively to the prerogative of system administrators should immediately be given a little advice: do not be alarmed, you can do the installation and configuration yourself. These processes are not as complicated as they might seem at first glance. But first, you should take into account some theoretical information.

Server operating system: what is it?

Let's start with the very essence of this software. Actually, operating systems of this type are used exclusively in local or virtual networks, when there is a certain number of child terminals there.

The server room can be installed in two ways, but it has one functional purpose - to ensure the operation of common applications on the network and the interaction of connected terminals with each other. Accordingly, we are talking about user groups that have access rights to a particular tool, resource or document.

Features of building server operating systems

In terms of installing a server-type OS, you should take into account the scheme of connecting computers into a single network. This is the so-called network topology.

The optimal option is the “star” scheme and derivatives based on it. Here, thick technology can be used and when the installation of the server operating system will be carried out exclusively on the central computer, which ensures the operation of all other machines and the corresponding software when loading the OS over the network, or there is a partial installation of child OS components on network terminals. That's not the point.

The main and most basic purpose of the server operating system is precisely to ensure maximum performance of all machines on the network and installed software, plus full control in terms of administering local terminals and the entire network as a whole.

Even using modern online games requires installing the OS on a central server. Many may disagree with this statement, believing that you can go to the game website from your home terminal and use all the features. This is wrong. The fact is that the home computer is still connected to the game server, and the load in terms of using the computing capabilities of the local computer falls only partially on it (the main operations are still performed on the central server, and the game client can be partially downloaded to the machine, with which the access attempt is made).

Conditional rating

The ratings of server operating systems below should not be considered absolutely accurate. The problem is that some operating systems are already outdated and are used very infrequently, others are difficult to configure. In general, there are a huge number of criteria according to which the priorities in the list can easily be changed

However, among all existing operating systems the most popular are the following:

Free BSD.
Windows Server (NT-based families of Windows server operating systems, 2003, 2008 R2, 2012 and higher).
CentOS.
Debian.
Red Hat Enterprise Linux.
Ubuntu Server.
Gentoo.
Fedora.
OS X Server.
OpebBSD.
Solaris.
HP-UX;.
AIX (IBM).
Netware (Novel).

It goes without saying that the first and last places are not established here, but only a list of the most popular server systems is given. Below, some of them will be discussed in a brief description of their capabilities, advantages and disadvantages.

Free BSD

Although this OS is a fairly popular system, it is nevertheless hopelessly outdated and losing global ratings in terms of use and performance.

The biggest problem is that too few software products have been developed for this OS in the form of commercial applications that could function on top of its platform. But an undoubted advantage is the ability to fine-tune the kernel and the availability of fairly powerful tools for working with memory, not to mention the input/output system.

CentOS

This server operating system is almost a complete analogue of Red Hat, but has expanded capabilities in terms of support.

Its advantage is that the system has a fairly fast package manager, as well as almost all hosting control panels.

Debian

This system is one of the branches of Linux. It is this OS that has become very widespread due to its versatility.

It is considered to be the most stable of all existing systems and in the installation package, along with KDE and GNOME, it also contains the office distribution kit LibreOffice. In addition, when installing this OS, you do not need to worry about installing additional audio and video codecs and decoders, since they are included in the package itself.

Red Hat Enterprise Linux

You can't deny the popularity of this system. It can be used as a reliable corporate OS with all the ensuing consequences.

It is not surprising that even the majority of the world's exchanges use this particular system to organize the work of their applications. Here we can also mention financial structures, telecommunications companies, and animation studios. The only negative with all the advantages is the fairly high cost.

Ubuntu

Another variety of Linux (or rather, UNIX-like systems), which has firmly established itself in the ranking.

Although its use is mostly associated with home computers, it is believed that this OS is best suited for installing a home server (if the user has limited financial resources). In principle, this system is very similar to Debian, however, unlike its ancestor, in this system the source code has been slightly reworked.

Gentoo

Before us is another modification of Linux. This platform is interesting because it is based on the GNU license.

There are no problems with stability. But, as noted by many experts, the system suffers a little in terms of security. Even if LAMP is used, security holes are found in the system.

Solaris

Solaris is considered to be one of the most powerful server operating systems ever developed.

What can I say? One has only to look at its capabilities. This OS supports about 576 GB of physical RAM, about 4 billion network connections, can support one million simultaneously running services and processes, and also has an open network environment. And that's not all...

OpenBSD

This system, according to most experts, is the most secure. Perhaps its operating speed is not as high as that of other modifications of Linux or UNIX, but if you learn how to use it, many will find it a very interesting solution, even in comparison with the most popular Windows systems.

Installation looks a little complicated, use and configuration too. But provided that the user is at least somewhat familiar with operating systems of this type, using the command line should not be particularly difficult.

Windows Server (2008 R2)

Finally, let's look at Windows server operating systems. Oddly enough, despite the popularity of desktop versions, these solutions are not particularly widespread in the world. The most basic problem is that Windows server operating systems, starting with the NT modification, do not have the necessary security features.

And this is recognized by all specialists in the field of administering systems of this type. For example, an antivirus for the Windows 7 server operating system must be installed. In connection with the latest hacker attacks, which were aimed mainly at computers using the seventh version of Windows (not to mention how long it took to install various kinds of updates that were supposed to prevent the virus from acting), the same built-in Defender service (or Microsoft Essentials) simply cannot cope with its responsibilities.

Administration of Windows server operating systems is, of course, terribly simplified. For example, you do not need to be an all-knowing system administrator to reassign IP, change or Proxy, establish a connection to a child terminal based on network download or by monitoring activity through an RDP client. Sometimes it is enough to even launch the built-in network connection diagnostic tool, after which a notification about a problem or error will be issued, if there are any in the system.

As for the rest, regarding troubleshooting tools, for example, on the server itself, you can use the command console with clearing the DNS cache or resetting the general IP configuration parameters.

For example, commands entered into the console might look like this:

ipconfig/flushdns,
ipconfig/registerdns,
ipconfig /renew
ipconfig/release.

Note: a command line with commands of this type works equally well to resolve problems on child user PCs and on servers.

What's the result?

As a main conclusion, we can only note that, alas, Windows systems and most of the developments created by Microsoft, in comparison with UNIX systems and the Linux variations that are based on them, are clearly inferior. The same FreeBSD OS is much more flexible in installation and configuration, not to mention the fact that it is most fully protected in the form of an open source modification that anyone can change or modify. Actually, the question itself arises: “What to prefer?” I don’t want to give any advice, but fundamentally, for organizing local networks using a central server, it is better to give preference to a UNIX-based OS (Linux) rather than Windows. It’s cheaper, easier, and safer. True, to users who are accustomed to the interface, this shell may seem, to put it mildly, not entirely clear. Nevertheless, understanding the interface will not be difficult. But you’ll have to tinker a little with the settings (what could we do without it?).

In turn, it should be noted that not all modern operating systems meet the requirements of today, although they have quite good potential for development and modernization. This primarily concerns FreeBSD, which, unfortunately, is not destined to develop, despite all the advantages. The developers simply put an end to it. But other systems are not lagging behind. So, in the near future we can even expect an updated version of Windows 10, which already combines the functionality of both the desktop and server models (after standard installation, the server can be configured quite simply). You don't even have to install additional clients. It is enough to use one of the components in the OS itself and in the BIOS settings (the same Hyper-V module for creating a virtual machine). With it, in terms of resource distribution between connected terminals, the server OS simply “flies”.

Installimage can be launched directly from the rescue system on your server. This is a user-friendly interface with the ability to select a Linux distribution. You get full control over the partitioning of your disk(s) and have the ability to activate software RAID or LVM using the user-friendly editor.

Autosetup

If installimage finds a file in the rescue system /autosetup, then this file is automatically used as a configuration file. If the file does not contain errors, the menu and editor are not displayed.

Using installimage

To activate the rescue system, go to our website.

Select Services – Products/services

In the line of the desired server on the right side, click “Product Management”

Press the “Rescue” softkey

In the field, select, for example, Linux x64 OS and click “Activate Rescue System”:

After activation, access data for Rescue Mode will appear. Copy the access data and then go to the “Reboot” field and reboot using one of three methods (if you do not reboot the server, Rescue Mode will not be activated). The fastest reboot method is “Send CTRL+ALT+DEL to the server,” but this method will not work if you already have Windows installed. To reboot in this case, you need to select “Execute a automatic hardware reset”.

Then you need to log in via SSH with the data you copied.

You can start the installation procedure:

root@ rescue ~ # installimage

You will see a menu where you can select your Linux distribution and image from the list in the submenu.

After selecting the image, you will receive a notification that the editor will launch, and the configuration file will open.

Midnight Commander (“mcedit”) is used as an editor.

The disks present in the system are displayed in the first line in the DRIVE variable<число>. Above each variable you can see the disk type. If you want to leave the disk unchanged in its current state, you can remove it from the configuration file by placing a “#” in front of it. Please note that the number in the DRIVE variable must be changed accordingly.

For example:

# ST31500341AS #DRIVE1 /dev/sda # ST31500341AS DRIVE1 /dev/sdb

If the server has several disks, then using the “SWRAID” and “SWRAIDLEVEL” variables you can create a software RAID array from them. Software RAID of all levels applies to all drives in the configuration (listed in the DRIVE variables). If you do not want to include a specific disk in the array, then you need to remove it from the configuration (as described above).

The “BOOTLOADER” variable allows you to select the bootloader to be installed; “Grub” or “Lilo” are available. When you select “grub”, depending on the operating system you are installing, GRUB2 or GRUB1 (legacy grub) will be installed.

The server name is set in the “HOSTNAME” variable

Installimage also supports setting up a disk partition scheme (including with LVM). The corresponding syntax is described in examples in the configuration file.

After exiting the editor using F10 (save and exit), the syntax of the configuration file will be checked. If the file contains errors, the editor will open again.

If you see this output after 1-15 minutes (depending on the image and partition), then your system is ready.

The root password is set to the current password of the rescue system.

Perform a “reboot” in the rescue system:

root@ rescue ~ # reboot

The newly installed system will boot and you can log into it, still using the rescue password.

Peculiarities

When using the installimage script to install Debian or Ubuntu, the time for cron jobs in the file /etc/cron.d/mdadm is given randomly.

Question answer

Am I losing data on my hard drive(s)?

Yes. All data on the selected disk(s) is completely erased and a new partition table is created.

Why can't I create a partition larger than 2 TB?

Creating partitions larger than 2 TB is only possible when using GUID Partition Tables (GPT). GRUB2 should be selected for this, as it supports booting from GPT disks.

The installation script throws one or many errors. What should I do now?

First, please restart the installation.

Do I need to put “all” at the end of the partition table or can I put this line not at the end?

The size parameter “all” in the configuration file means use all currently available hard drive space. Since partitions are created one after another, after specifying “all” the table will end, since there will be no more free space. Of course, you don’t have to use “all” at all.

I have two (or more) disks installed. Can the installation script automatically create software RAID?

Yes, it can create software RAID levels 0, 1, 5, 6 and 10. There are two options in the configuration file for this: SWRAID And SWRAIDLEVEL. If option SWRAID activated (1), software RAID level is created SWRAIDLEVEL.

Pressing the F10 key does not work. After pressing, “~21” (or something similar) is displayed.

Press Escape and then “0”. In most cases this replaces F10.

Can I use logical partitions ( LVM)?

Yes.

I get the error “Due to maintenance the installimage-script is temporarily unavailable. We are sorry for the inconvenience.”

If you see this message, it means parts of the script are being updated. This usually takes a couple of minutes. To avoid errors, blocking is required while updates are installed. You will be able to use the script within a couple of minutes after this message appears.

Who is the author of the script? Can I use it freely?

The script is publicly available and the programming language is available under GPL. The site company is not responsible for any damage caused by modified versions of the script. There is no support provided for any issues that involve script modification.

What is the root user password for MySQL in a LAMP image?

The root user password for MySQL when installing a LAMP image after a fresh installation is located in the /password.txt file.

Liked? We tried, really! Do us a favor - subscribe to new articles. Next is better!

Related materials:

Site Map