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

How to build a private Linux mini system, in view of this problem, this article introduces the corresponding analysis and answer in detail, hoping to help more partners who want to solve this problem to find a more simple and feasible way.

I. Preface

Since the birth of the Linux operating system on October 5, 1991, its openness and freedom have been favored by many technology cattle, and each Linux enthusiast has contributed his part to it, whether in the Linux kernel or open source software, etc., all provide a good learning and research environment for our later generations. As a Linuxer, thank you for providing us with a free space, so that we can also study Linux while learning.

The following is mainly by tailoring the existing Linux system to create a small Linux system of its own, so that it can load the network card driver and configure the IP address to achieve network functions.

Second, principle

Introduction to the startup process:

Before making the Linux mini system, it is necessary to take a look at the startup process of Linux:

1. First of all, Linux has to pass the POST self-test to check whether the hardware is faulty.

2. If you have multiple boot disks, you need to select the boot disk in BIOS

3. Start the bootloader bootstrap in MBR

4. Load kernel files

5. Init, the parent process and ancestor of all processes

6. Print welcome interface

In the startup process of Linux, two other files are needed to load the kernel file:

1) initrd, which is a disk device simulated in memory on CentOS5

2) initramfs, which is a file system simulated in memory on CentOS6

What is the main purpose of init in the departure process?

Init calls the configuration file / etc/inittab, and then executes the system initialization script of / etc/rc.d/rc.sysinit

Enlighten

After printing the welcome interface of Linux, it shows that the system has started successfully. If we want to make a small Linux system, we only need to load all the files used in the boot process together to light up our own system. While Linux is a modular operating system, many functional components are implemented through modular tools and support dynamic loading and unloading. If we want to achieve a certain function, we only need to load the corresponding module, and we can greatly reduce the size of our Linux operating system.

III. Operation steps

1. Target disk partition

Hang a new disk on the host, named soft-Linux, this disk is the second disk on the host, so here is / dev/sdb, and when it is mounted to the target host, because there is only one disk there, the name on the target host should be / dev/sda, this should not be confused. First, we need to divide the target disk into two areas and format them. * 500m partition is used to mount the bootstrap program, and the second partition 10G is used to mount the root file system. Then mount / dev/sdb1 to / mnt/boot and / dev/sdb2 to / mnt/sysroot.

[root@nmshuishui ~] # mount / dev/sdb1 / mnt/boot mount: mount point / mnt/boot does not exist [root@nmshuishui ~] # mkdir-p / mnt/boot / mnt/sysroot [root@nmshuishui ~] # mount / dev/sdb1 / mnt/boot [root@nmshuishui] # mount / dev/sdb2 / mnt/sysroot/ [root@nmshuishui ~] #

2. Install grub to the target disk

A system can boot, so we need to install a grub bootstrap program to our new disk first. There are two main commands to install grub bootstrap program, one is grub-install, the other is setup, here * use grub-install to install. Because:

① grub-install will install the grub boot second phase file ② setup will not install the second stage boot program, is the installation boot information to MBR the second need to pay attention to is-root-directory= followed by the path should be the boot directory, not / mnt/boot, because the boot directory is under mnt; the target disk is / dev/sdb [root@nmshuishui ~] # grub-install-- root-directory=/mnt / dev/sdb Probing devices to guess BIOS drives. This may take a long time. Installation finished. No error reported. This is the contents of the device map / mnt/boot/grub/device.map. Check if this is correct or not. If any of the lines is incorrect, fix it and re-run the script `grub-install'. (fd0) / dev/fd0 (hd0) / dev/sda (hd1) / dev/sdb [root@nmshuishui ~] # cd / mnt/boot/ [root@nmshuishui boot] # ls grub lost+found [root@nmshuishui boot] # cd grub/ [root@nmshuishui grub] # ls device.map e2fs_stage1_5 fat_stage1_5 ffs_stage1_5 iso9660_stage1_5 jfs_stage1_5 minix_stage1_5 reiserfs_stage1_5 stage1 stage2 ufs2_stage1_5 vstafs_stage1_5 xfs_stage1_5 [root@nmshuishui grub] #

After installing grub, enter the grub directory, we will find that there is no grub.conf configuration file, so our bootstrap program is not sound, so we need to write a configuration file in it manually, but we need to know the kernel version, and then go back to supplement this step after porting the kernel version.

3. Copy kernel files and initrd files

Init is the program used to generate all other processes in the system. It exists as a daemon, and its process number is 1, which is the parent and ancestor of all processes, so it is impossible not to migrate. It invokes the configuration file / etc/inittab and then executes the system initialization script for / etc/rc.d/rc.sysinit.

Copy the kernel file and initrd file to the boot directory under / dev/sdb.

[root@nmshuishui grub] # cp / boot/vmlinuz-2.6.32-358.el6.x86_64 / mnt/boot/vmlinuz-soft [root@nmshuishui grub] # cp / boot/initramfs-2.6.32-358.el6.x86_64.img / mnt/boot/initramfs-soft.img [root@nmshuishui grub] #

4. Create the root file system of the target host

① creates a file system using command line deployment

[root@nmshuishui sysroot] # mkdir-pv / mnt/sysroot/ {etc/rc.d,usr,var,proc,sys,dev,lib,lib64,bin,sbin,boot,srv,mnt,media,home Root} mkdir: created directory `/ mnt/sysroot/etc' mkdir: created directory` / mnt/sysroot/etc/rc.d' mkdir: created directory `/ mnt/sysroot/usr' mkdir: created directory` / mnt/sysroot/var' mkdir: created directory `/ mnt/sysroot/proc' mkdir: created directory` / mnt/sysroot/sys' mkdir: created directory `/ mnt/sysroot/dev' mkdir: created directory` / mnt/sysroot/lib' mkdir: created directory `/ mnt/sysroot/lib64' mkdir: created directory` / Mnt/sysroot/bin' mkdir: created directory `/ mnt/sysroot/sbin' mkdir: created directory` / mnt/sysroot/boot' mkdir: created directory `/ mnt/sysroot/srv' mkdir: created directory` / mnt/sysroot/mnt' mkdir: created directory `/ mnt/sysroot/media' mkdir: created directory` / mnt/sysroot/home' mkdir: created directory `/ mnt/sysroot/root' [root@nmshuishui sysroot] # ls bin boot dev etc home lib lib64 lost+found media mnt proc root sbin srv sys usr var [root@nmshuishui sysroot] #

② migrates bash commands and their library files to the root file system

[root@nmshuishui mnt] # sh ~ / scripts/cporder.sh Enter a command: bash Enter a command: shutdown Enter a command: vim Enter a command: touch Enter a command: mkdir Enter a command: rm Enter a command: ls Enter a command: cat Enter a command: less Enter a command: ifconfig Enter a command: ip Enter a command: route Enter a command: quit quit [root@nmshuishui mnt] # sync [root@nmshuishui mnt] # sync [root@nmshuishui mnt] # ls boot sysroot [root@nmshuishui mnt] # cd Sysroot/ [root@nmshuishui sysroot] # ls bin lib64 sbin usr [root@nmshuishui sysroot] # cd bin/ [root@nmshuishui bin] # ls bash cat ls mkdir rm touch [root@nmshuishui bin] # ln-sv bash sh `sh'-> `bash' [root@nmshuishui bin] # sync [root@nmshuishui bin] #

Attached: command migration script

#! / bin/bash # target=/mnt/sysroot clearCmd () {if which $cmd & > / dev/null Then cmdPath= `which-- skip-alias $cmd` else echo "No such command" return 5 fi} cmdCopy () {cmdDir= `dirname $1` [- d ${target} ${cmdDir}] | | mkdir-p ${target} ${cmdDir} [- f ${target} ${1}] | | cp $1 ${target} ${cmdDir} libCopy () {for lib in `ldd $1 | grep-o "/ [^ [: space:]]\ {1,\}" ` Do libDir= `dirname $lib` [- d ${target} ${libDir}] | | mkdir-p ${target} ${libDir} [- f ${target} ${lib}] | | cp $lib ${target} ${libDir} done} while true; do read-p "Enter a command:" cmd if ["$cmd" = = 'quit']; then echo "quit" exit 0 fi clearCmd $cmd [$?-eq 5] & & continue cmdCopy $cmdPath libCopy $cmdPath done

5. Provide configuration files for grub

With the kernel and initrd files migrated above, we can write the grub.conf configuration file based on the kernel version and initrd version:

[root@nmshuishui grub] # vim grub.conf default=0 timeout=5 title nmshuishui soft-Linux root (hd0,0) kernel / vmlinuz-soft ro root=/dev/sda2 quiet selinux=0 init=/bin/bash initrd / initramfs-soft.img ~ quiet is a silent installation and no longer displays a lot of information at the time of installation. Later, turn off selinux, and init will use / bin/bash, telling the kernel not to look for init programs. If you do not specify this step, you will report kernel panic (kernel panic) during startup, thinking that the system is the only one, and you can't panic without the init process.

6. Start the test

7. Special reminder

If you do this experiment on vmvare, when creating a new virtual machine to create a new disk, be sure to select "Store virtual disk as a single file", otherwise, there will be kernel panic kennel panic.

Fourth, load the module to realize the network function

1. View the Nic module information of the host.

[root@nmshuishui net] # lsmod | grep e1000e1000 170646 0 [root@nmshuishui net] #

2. View the details of the network card

[root@nmshuishui net] # modinfo e1000 filename: / lib/modules/2.6.32-358.el6.x86_64/kernel/drivers/net/e1000/e1000.ko version: 7.3.21-k8-NAPI license: GPL description: Intel (R) PRO/1000 Network Driver author: Intel Corporation Srcversion: 1D4F1E82BB99EA36D320B1B alias: pci:v00008086d00002E6Esv*sd*bc*sc*i* alias: pci:v00008086d000010B5sv*sd*bc*sc*i* alias: pci:v00008086d00001099sv*sd*bc*sc*i* alias: pci:v00008086d0000108Asv*sd*bc*sc*i* alias: pci:v00008086d0000107Csv*sd*bc*sc*i*

Here, the path of the Nic module is queried and copied to the library file of / dev/sdb:

[root@nmshuishui net] # mkdir-pv / mnt/sysroot/lib64/modules mkdir: created directory `/ mnt/sysroot/lib64/modules' [root@nmshuishui net] # cp / lib/modules/2.6.32-358.el6.x86_64/kernel/drivers/net/e1000/e1000.ko / mnt/sysroot/lib64/modules/e1000.ko

3. Init program

Now, although the module has been copied over, it cannot be used, and now it does not meet our process needs, because there is not even the most basic init program. If we want this init, we have two options: *, transplant the host system, but the format will be more complicated; so we'd better write our own script first, using the script as init, which can make the small system run. Init is usually in the sbin directory, so we'll write an init script on the / dev/sdb2 partition.

[root@nmshuishui ~] # cd / mnt/sysroot/sbin/ [root@nmshuishui sbin] # vim init #! / bin/bash # print Welcome info echo-e "Welcome to\ 033 [34m nmshuishui soft-Linux\ 033 [0m" # mount wei wenjian system when the system is running. Mount-n-t proc proc / proc mount-n-t sysfs sysfs / sys # mount ethernet driver autl when the system is start. Insmod / lib64/modules/e1000.ko [$?-eq 0] & & echo-e "Load e1000 module succeeded [\ 033 [32m0K\ 033 [0m]" ifconfig lo 172.0.0.1 ifconfig lo 8 ifconfig eth0 172.16.251.235 mount the 16 # mount the / dev/sda2 to make it can be write and read. Mount-n-o remount,rw / dev/sda2 / # run / bin/bash / bin/bash

After writing this init script, we want to give it an execute permission so that it can be executed; mount,insmod commands are also used in this script, so we need to use the last script to port these commands. * I also need to replace the init=/bin/bash in / mnt/boot/grub/grub.conf with init=/sbin/init, because now I'm going to use this init script to perform system startup, and I don't need / bin/bash to replace it anymore.

4. Linux mini system to realize the network function.

After the above steps are completed, you can hang / dev/sdb on another host to experience our private customized applet.

This is the answer to the question on how to build a private Linux mini system. 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.