Article Directory
One, Linux operating system boot process
System boot is the beginning of the operation of the operating system. Before the user can log in to the system normally, the Linux boot process completes a series of initialization tasks and loads the necessary programs and command terminals to prepare for the user to log in.
1. Overview of the boot process
1.1 Power-on self-test
- After the server host is turned on, the CPU, memory, graphics card, keyboard and other devices will be preliminarily tested according to the settings in the motherboard BIOS (Basic Input Output System). After the test is successful, the system control will be transferred according to the preset startup sequence. Most of the time, it will be transferred. To the local hard drive
- Detect the first device capable of booting the system, such as a hard disk or optical drive
1.2 MBR boot
- When starting the system from the local hard disk, first transfer system control to the partition containing the operating system boot file according to the MBR (Master Boot Record) setting in the first sector of the hard disk; or directly according to the boot in the MBR record Information to call the boot menu (eg GRUB)
- Run the boot GRUB boot program placed in the MBR sector
1.3 GRUB menu
- For the Linux operating system, GRUB (Unified Boot Loader) is the most widely used multi-system bootloader program. After the
system control is passed to GROB, the boot menu will be displayed for the user to choose, and according to the selected option (or use the default value) ) Load the Linux kernel file, and then transfer system control to the kernel.
Centos 7 uses GRUB2 boot loader.
GRUB boot program reads the GRUB configuration file /boot/grub2/grub.cfg to obtain the kernel and mirror file system settings
1.4 Load the Linux kernel
- The Linux kernel is a pre-compiled special binary file, between various hardware resources and system programs, responsible for resource allocation and scheduling. After the
kernel takes over the control of the system, it will fully control the running process of the entire Linux operating system - In Centos 7 system, the default kernel file is located at "/boot/vmlinuz-3.10.0-514.e17.x86_64"
- Load the kernel and mirror file system into memory
1.5 init process initialization
- In order to complete the further system boot process, the Linux kernel first loads the "/sbin/init" program in the system into the memory to run (the running program is called a process), the init process is responsible for completing the initialization of the entire system, and finally waits for the user to proceed log in
- Load the hardware driver, the kernel loads the init process into the memory to run
2.init process
- The processes (running programs) in the Linux system are marked with numbers. The identification number of each process is called PID. In the process of booting Linux, "/sbin/init" is the first program loaded by the kernel. Therefore, the PID number corresponding to the init process is always "1"
- After the init process runs, it will continue to execute other programs in the system, and continuously generate new processes. These processes are called child processes of the init process. On the other hand, the init process is the parent process of these processes.
Of course, these child processes can also be further generated. The respective sub-processes continue to multiply at a time, and finally form a leafy process tree, which together provide services for users - Init is the "ancestor" of all processes that maintain the entire Linux system. Therefore, the init process is not allowed to be easily terminated. When you
need to switch between different system operating states, you can send the correct execution parameters to the init process, which is done by init itself. Related operations
3.Systemd
- Systemd is an init software of Linux operating system
- CentOS7 uses a new Systemd startup method to replace the traditional SysVinit
- Traditional SysVinit relies on the serial execution of Shell scripts to start the service, resulting in low efficiency and slow system startup
- Systemd can start more service processes in parallel, and has the ability to start services on demand, so that fewer processes are started, thereby increasing the system startup speed
- The first init process running in CentOS7 is /lib/systemd/systemd
4. Systemd unit type
| Unit type | extension name | Description |
|---|---|---|
| Service | .service | Describe a system service |
| Socket | .socket | Describe a socket for inter-process communication |
| Device | .device | Describe a device file recognized by the kernel |
| Mount | .mount | Describe the mount point of a file system |
| Automount | .automount | Describe a memory swap device or swap file |
| Swap | .swap | Describe a memory swap device or swap file |
| Path | .path | Describe a file or directory in a file system |
| Timer | .timer | Description-a timer (used to implement cron-like scheduling tasks) |
| Snapshot | .snapshot | Used to save the state of a systemd |
| Scope | .scope | Use systemd's bus interface to programmatically create external processes |
| Slice | .slice | Describe a group of management system processes organized by hierarchy in Cgroup |
| Target | .target | Describe a set of systemd units |
5. Systemd target corresponding to the run level
| Run level | Systemd的target | Description |
|---|---|---|
| 0 | target | Shutdown state, the host will be shut down when using this level |
| 1 | rescue.target | Single user mode, you can log in to the system without password verification, mostly used for system maintenance |
| 2 | multi-user.target | User-defined/domain specific run level. The default is equal to 3 (access to the network is not supported) |
| 3 | multi-user.target | Full multi-user mode with character interface, most server hosts run at this level |
| 4 | multi-user.target | User-defined/domain specific run level. The default is equivalent to 3 |
| 5 | graphical.target | The multi-user mode of the graphical interface provides a graphical desktop operating environment |
| 6 | reboot.target | Restart, the host will be restarted when using this level |
2. Eliminate startup faults
1. Repair MBR sector failure
1.1 Causes of failure
- Damage caused by viruses, Trojan horses, etc.
- Incorrect partition operation, disk read and write errors
1.2 Failure phenomenon
- Cannot find the boot program, start interrupted
- Unable to load operating system
1.3 Solutions
- Backup files should be made in advance
- Boot into emergency mode with installation CD
- Restore from backup file
1.4 Practice
1.4.1 Back up MBR sector data to other disks (/dev/sdb1)
- First add a new hard drive
- New partition
[root@localhost ~]# fdisk /dev/sdb
欢迎使用 fdisk (util-linux 2.23.2)。
更改将停留在内存中，直到您决定将更改写入磁盘。
使用写入命令前请三思。
Device does not contain a recognized partition table
使用磁盘标识符 0xbd9ad88f 创建新的 DOS 磁盘标签。
命令(输入 m 获取帮助)：n
Partition type:
p primary (0 primary, 0 extended, 4 free)
e extended
Select (default p): p
分区号 (1-4，默认 1)：
起始 扇区 (2048-83886079，默认为 2048)：
将使用默认值 2048
Last 扇区, +扇区 or +size{K,M,G} (2048-83886079，默认为 83886079)：
将使用默认值 83886079
分区 1 已设置为 Linux 类型，大小设为 40 GiB
命令(输入 m 获取帮助)：w
The partition table has been altered!
Calling ioctl() to re-read partition table.
正在同步磁盘。
- format
[root@localhost ~]# mkfs.xfs /dev/sdb1
meta-data=/dev/sdb1 isize=512 agcount=4, agsize=2621376 blks
= sectsz=512 attr=2, projid32bit=1
= crc=1 finobt=0, sparse=0
data = bsize=4096 blocks=10485504, imaxpct=25
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0 ftype=1
log =internal log bsize=4096 blocks=5119, version=2
= sectsz=512 sunit=0 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0
- Mount
[root@localhost ~]# mkdir /backup
[root@localhost ~]# mount /dev/sdb1 /backup/
- Backup
[root@localhost ~]# cd /backup/
[root@localhost backup]# ls
[root@localhost backup]# dd if=/dev/sda of=/backup/mbr.bak bs=512 count=1
记录了1+0 的读入
记录了1+0 的写出
512字节(512 B)已复制，0.000197902 秒，2.6 MB/秒
[root@localhost backup]# ls
mbr.bak
1.4.2 Simulating MBR boot sector destruction
- Simulated destruction
[root@localhost backup]# dd if=/dev/zero of=/dev/sda bs=512 count=1
记录了1+0 的读入
记录了1+0 的写出
512字节(512 B)已复制，9.7308e-05 秒，5.3 MB/秒
- Restart the system and test it
[root@localhost backup]# init 6
Connection closing...Socket close.
Connection closed by foreign host.
Disconnected from remote host(CentOS 7-3) at 16:58:38.
Type `help' to learn how to use Xshell prompt.
[D:\~]$
1.4.3 Boot interface
- After restarting, I found that the system interface was not entered, but the installation wizard appeared
- Here we select the "Troubleshooting" option to enter the emergency mode
- Type "1" to select Continue and press Enter to continue
- Press Enter again to enter the Bash Shell environment with "sh-4.2#" prompt
sh-4.2# mkdir /backupdir
sh-4.2# mount /dev/sdb1 /backupdir ##挂载带有备份文件的分区
sh-4.2# dd if=/backupdir/mbr.bak of=/dev/sda bs=512 count=1 ##恢复备份数据
sh-4.2# exit ##执行exit命令退出临时Shell环境，系统将会自动重启
- OK The system starts to restart, and then it is found to be normal, you can directly enter the system
2. Repair GRUB boot failure
2.1 Causes of failure
- The GRUB boot program in MBR is corrupted
- The grub.conf file is missing, the boot configuration is incorrect
2.2 Failure phenomenon
- System booting stalls, displaying "grub>" prompt
2.3 Solutions
- Try to manually enter the boot command
- Enter emergency mode, rewrite or restore grub.conf from backup
- Rebuild the grub program into the MBR sector
2.4 Practice
- The boot interface enters emergency mode, and the GRUB menu configuration file is rebuilt
2.4.1 Simulated failure
/boot/grub/目录下的包是用于启动菜单的背景图片及样式
/boot/grub2/grub.cfg ##Grub配置文件
- Delete grub.cfg to simulate failure
- Here we do not need to backup, because it is directly rebuilt
[root@localhost boot]# cd grub2
[root@localhost grub2]# ls
device.map fonts grub.cfg grubenv i386-pc locale
[root@localhost grub2]# rm -rf grub.cfg
[root@localhost grub2]# ls
device.map fonts grubenv i386-pc locale
- Enter reboot to restart, then you will enter this interface
- Attention, how do we enter the boot interface at this time, the hand speed must be fast, press "ESC" immediately after clicking the restart, try several times, it will always succeed
- Here we choose CD
2.4.2 Boot interface
- Enter emergency mode and restore CenOS system
- Enter 1, press enter
- Switch to the system root environment (this time you can use the TAB key to complete)
sh-4.2# chroot /mnt/sysimage
bash-4.2#
- Reinstall the GRUB boot program to the MRB sector of the first hard disk (/dev/sda)
bash-4.2# grub2-install /dev/sda
- Rebuild the configuration file of the GRUB menu
bash-4.2# grub2-mkconfig -o /boot/grub2/grub.cfg
- Exit the chroot environment and restart
bash-4.2# exit
sh-4.2# reboot
- OK, you can enter the system interface, let's verify it again and find that grub.cfg has it again
[root@localhost ~]# cd /boot/grub2
[root@localhost grub2]# ls
device.map fonts grub.cfg grubenv i386-pc locale