What is a process ？
process （Process） It is an execution activity of a program in its own virtual address space . The reason to create a process , Just to make
Multiple programs can be executed concurrently , So as to improve the resource utilization and throughput of the system . Simply put, a process is the execution of a program .
What is the difference between a process and a program ？
Program ： Just a static set of instructions ; A process is a dynamic execution process of a program , It has a life cycle , Is the dynamic production and extinction .
process ： Is a resource request 、 Dispatch and independent operation unit , therefore , It uses running resources in the system ; And the program cannot apply for system resources 、 Cannot be scheduled by the system 、
Nor can it be used as an independent unit , therefore , It does not occupy the operating resources of the system .
Neither program nor process One correspondence . On the one hand, a program can be shared by multiple processes , That is, when a program is running, multiple processes can be generated ; On the other hand , A process can execute several programs in sequence during its lifetime .
Properties of the process
stay linux There are always many processes running in the system at the same time , Each process has an identification number , be called PID（Process ID）, To distinguish different processes . Each process has a ID Number , Can pass ps The order shows
It can be used echo $$, see shell Hezi shell
[[email protected] ~]# echo $$ first $ On behalf of the variable , the second $ A variable's value , this $ On behalf of the current shell The process number of
2943
[[email protected]~]# bash Intron shell
[[email protected]~]# echo $$
2983 The process number is different from that of the parent process
View the system process tree
[[email protected]~]# pstree -p
systemd(1)─┬─ModemManager(909)─┬─{ModemManager}(938)
│ └─{ModemManager}(964)
systemd It is the first process that the system starts , Is the parent of all processes , Its process number is 1.
Usually killing a child process does not affect the parent process , Killing the parent process will affect the child process , and systemd This process cannot be killed , Unless shut down or restart .
Hard disk 、 process 、 Memory 、CPU The relationship between
One of the three illusions of life ：QQ、 Storm video 、 Browsers, etc. can be executed together . Are processes executed in parallel . No more than that. . For a single CPU For the system, at a certain moment , Only one process can be in execution , occupy CPU Right to use , Other processes are in other States , Wait for system resources , Each process constantly switches between certain states according to the scheduling algorithm . But because CPU Faster processing efficiency , So that users can't feel the changes .
State of process
R (TASK_RUNNING), Executable state .
S (TASK_INTERRUPTIBLE), An interruptible sleep state
D (TASK_UNINTERRUPTIBLE), Uninterrupted sleep
T (TASK_STOPPED or TASK_TRACED), Pause state or trace state .
Z (TASK_DEAD - EXIT_ZOMBIE), Exit state , The process becomes a zombie process .
X (TASK_DEAD - EXIT_DEAD), Exit state , The process is about to be destroyed .
R (TASK_RUNNING), Executable state .
Running or ready ： Ready status means that the process has applied for CPU All other resources except .
experiment ：
Open Firefox browser and use top Command view status
S (TASK_INTERRUPTIBLE), An interruptible sleep state Light sleep
In the waiting line , Wake up when resources are available （ For example, wait for keyboard input 、socket Connect 、 Signals and so on
experiment ：
Open Firefox browser and use top Command view status
D (TASK_UNINTERRUPTIBLE), Uninterrupted sleep Deep sleep
In the waiting line , Wake up when resources are available （ For example, wait for keyboard input 、socket Connect 、 Signals and so on ）, But it cannot be awakened by interruption .
experiment ：
Open Firefox browser and use top Command view status
T (TASK_STOPPED or TASK_TRACED), Pause state or trace state .
The process was suspended by an external program （ If received SIGSTOP The signal , The process goes into TASK_STOPPED state ）, Continue to run when agreed again （ Process received SIGCONT The signal , Get into TASK_RUNNING state ）.
Pause jobs Mission
Z (TASK_DEAD - EXIT_ZOMBIE), Exit state , The process becomes a zombie process . A dead state . Process resource user space is freed , But processes in the kernel PCB No release . Wait for the parent process to recycle
X (TASK_DEAD - EXIT_DEAD), Exit state , The process is about to be destroyed . The normal exit
Process scheduling process 
Check the process
The goal of learning process management ：
1、 View the current process of the system , Is there a process that uses too much resources
2、 Let the program execute first by adjusting the process priority
3、 How to deal with crashed programs stuck in memory
Common commands
PS： List process information statically
top： List process information dynamically
PS：
ps -l Check yourself bash Program
vim file1 &
ps -l
F： Represents this program tag
4: Indicates that the program permission is root
1: Indicates that the subroutine is only copied fork Without actually executing exec
S： Represents the state of the program STAT, The process status is 5 Kind of .
R （running）： The program is running ：
S（sleep）： The program is in sleep idle, Can be awakened （signal）
D： A sleep state that cannot be awakened , Usually this program may be waiting I／O.
T： Stop state （stop）, May be jobs A background program or traced Tracking status
Z（Zombie）： Zombie state , The program has terminated but cannot be removed from memory .
UID／PID／PPID： owner ／ process ID／ The parent process of this program ID
C：cpu Usage rate , The unit is percentage
PRI／NI ：priority／Nice Abbreviation , The representative program is cpu Priority of execution , The smaller it is , The first to be executed
ADDR／SZ／WCHAN：ADDR Memory related , If the status is a running program running, General display - .SZ Represents how much memory the program uses .WCHAN Indicates whether the program is running , If it is to use - Express .
TTY： Login terminal location
TIME： cost cpu Running program time
CMD：
ps -ef List all processes and trust information Than ps -l detailed
ps -A Show all processes
ps -e And the effect of this parameter is specified "A" Parameters are the same .
f Display in the form of a process tree
[[email protected]~]$ ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 22:39 ? 00:00:03 /usr/lib/systemd/systemd --switche
root 2 0 0 22:39 ? 00:00:00 [kthreadd]
ps aux Check all system running programs
a Display all programs under the current terminal , Including other users' programs .
u Display program status in a user oriented format .
x Show all programs , Do not differentiate by terminal .
yum repolist all
yum install -y httpd
systemctl start httpd
systemctl status httpd
ps aux
ps aux | grep httpd
top：
1、 Command options ：
-d: Screen refresh interval top -d second
2、 Shortcut keys in commands ：
h Show shortcut help
k terminate a process Process number 9
q Exit procedure
r Rearrange a priority for progress
s Change refresh interval , Unit second
l Switch to display the average load and start-up time information
m Toggle display memory information
t Toggle display progress and CPU State information
c Toggle display command name and full command line
M Sort by memory usage
P according to CPU Sort usage （ Default sort ）
T According to time / Sort in progressive time
1 Unfold multicore cpu Show
3、top Content description
top- Current system time
How long has the system been powered on
Current number of users
load average cpu Average load , The three values are ,1 minute ,5 minute ,15 minute
Current process number of task system , A total of ： The total number of processes , function ： Number of running processes , sleep ： The number of sleep processes , Has stopped ： Number of processes stopped , Corpse ： Number of zombie processes
％Cpu（s）cpu Usage rate , We ： The user to use cpu One hundred percent ,sy： The system kernel uses cpu One hundred percent ,id： remainder cpu One hundred percent
Mem Memory usage information , A total of ： Total memory size , You can use ： Cached memory , Already used ： Used memory ,buff / cache： The size of the cache memory
Exchange virtual memory information
PID process ID
User process owner
PR priority
NI nice value , Negative value indicates high priority , A positive value indicates a low priority
VIRT The amount of virtual memory used by the process
RES The amount of physical memory used by the process
SHR Shared memory size
S Process status ,D： Uninterrupted sleep ,R： function ,S： sleep ,T： track / stop it ,Z： Zombie process
％CPU Used by process CPU Percentage occupied
％MEM Percentage of physical memory used by the process
TIME + Progress used CPU Total time
COMMAND Command name
Kill process ：
kill Signal number PID
-1： Reload
-9： Force to kill
-15： The normal exit
[[email protected] mnt]# ps aux | grep vim
root 7258 0.0 0.4 33972 8152 pts/1 T 22:18 0:00 vim file
root 7490 0.0 0.0 12320 1076 pts/0 S+ 22:32 0:00 grep --color=auto vim
[[email protected] mnt]# kill -9 7258
[[email protected] mnt]# ps aux | grep vim
root 7504 0.0 0.0 12320 1056 pts/0 R+ 22:32 0:00 grep --color=auto vim
killall Process name
[[email protected] mnt]# ps aux | grep httpd
root 6424 0.0 0.6 273848 10816 ? Ss 21:47 0:00 /usr/sbin/httpd -DFOREGROUND
apache 6426 0.0 0.4 286064 8312 ? S 21:47 0:00 /usr/sbin/httpd -DFOREGROUND
apache 6429 0.0 0.5 1474980 9652 ? Sl 21:47 0:00 /usr/sbin/httpd -DFOREGROUND
apache 6430 0.0 0.5 1343852 9652 ? Sl 21:47 0:00 /usr/sbin/httpd -DFOREGROUND
apache 6431 0.0 0.5 1343852 9652 ? Sl 21:47 0:00 /usr/sbin/httpd -DFOREGROUND
root 7544 0.0 0.0 12320 1028 pts/0 R+ 22:32 0:00 grep --color=auto httpd
[[email protected] mnt]# killall httpd
[[email protected] mnt]# ps aux | grep httpd
root 7570 0.0 0.0 12320 1072 pts/0 R+ 22:34 0:00 grep --color=auto httpd
[[email protected] mnt]# pgrep vim | xargs kill -9
remarks ：
Linux Medium buff/cache Memory
We use it free、top And other related commands that can query the current memory usage , There is always one buff/cache Let's be confused .
buffer
When writing to disk , Save to disk buffer first (buffer), Then write to disk .
cache
That is, when reading the disk , After the data is read from the disk , Stay in the buffer (cache), Prepare for the use of subsequent procedures .
How to divide buffer/cache（ Already used or Free ）
Operating system perspective ： This memory is indeed used by the operating system .
User perspective ： If the user wants to use , This memory can be quickly recycled and used by user space programs , Therefore, from the user's point of view, this memory should be classified as idle state .
Linux The benefits of this mechanism
This is a Linux A very excellent design , The purpose is to upgrade the disk IO Performance of , Data read from low-speed block devices is temporarily stored in memory , Even though the data was no longer needed at the time , But the next time the application accesses that data , It can be read directly from memory , Bypassing low-speed block devices , So as to improve the overall performance of the system .
In order to improve system performance and not waste memory ,linux Make more memory cache, In order to improve the io Speed . Your memory is not occupied .
Linux Work scheduling jobs
because linux It is a multi person and multi task operating system , So users are using linux There will be some work that we need to watch the progress of completion , And we can just put some work directly in the background , Here we are concerned with the front and back stage implementation of the task , So how to put a task into the background for execution ?
[[email protected] ~]# firefox & #& The symbol is to put the process into the background for execution
[1] 9283
[[email protected] ~]# jobs # View background tasks
[1]+ Running firefox &
Conclusion , The program runs in the background , Does not affect the command prompt
Work management job control
adopt job control In a bash Perform multiple tasks in the environment .
We can't put tty1 The task of tty2 Execute below .
Suppose we have only one terminal interface . Therefore, the command prompt interface can appear, which we call foreground foreground, As for other work, you can put it in the background background Take a pause or work . Work put into the background, such as vim It can't be carried out , You can't use ctrl+c To terminate .
Something to watch out for ：
These work starting procedures must come from you shell The subroutine of （ Only manage your own bash）;
prospects ： The environment in which you can control and give orders is called foreground work （foreground）;
background ： Work that can operate on its own , You can't use ctrl+c Stop it , have access to bg／fg Call the job ;
In the background ：‘ perform ’ Your program can't wait terminal／shell Output input
Example ：
1、tar -zpcf /tmp/etc.tar.gz /etc &
[[email protected] desktop ]# tar -zpcf /tmp/etc.tar.gz /etc &
[1] 10502 1 Is the job number ,10502 Is a process,
[1]+ Done tar -zpcf /tmp/etc.tar.gz /etc appear Done Represents this task to complete
Put the tasks in the work into the background jobs
Execute two orders
vim ～／.bashrc After implementation ctrl+z
find ／ print Press after execution ctrl+z
jobs View the tasks in the background
Options ：
-l Show PID
-r Display in progress running The task of
-s Show stop The task of
fg % Task number Transfer the backstage to the front desk
bg % Task number Let the tasks in the background run in the background
kill -9 % Task number
[[email protected] ~]# dd if=/dev/zero of=/bigfile bs=1M count=4000 # Press ctrl+z Manually cut into the background
^Z
[4]+ Has stopped dd if=/dev/zero of=/bigfile bs=1M count=4000
[[email protected] ~]# jobs
[1] Running firefox &
[2] Has stopped vim 1.sh
[3]- Has stopped vim 2.sh
[4]+ Has stopped dd if=/dev/zero of=/bigfile bs=1M count=4000 # The task of manual entry will be suspended
[[email protected] ~]# jobs -l
[1] 9283 Running firefox &
[2] 9689 stop it (tty Output ) vim 1.sh
[3]- 9808 stop it vim 2.sh
[4]+ 9839 stop it dd if=/dev/zero of=/bigfile bs=1M count=4000
[[email protected] ~]# jobs -r
[1] Running firefox &
[[email protected] ~]# jobs -s
[2] Has stopped vim 1.sh
[3]- Has stopped vim 2.sh
[4]+ Has stopped dd if=/dev/zero of=/bigfile bs=1M count=4000
[[email protected] ~]# bg %4 # The background task continues to run
[4]+ dd if=/dev/zero of=/bigfile bs=1M count=4000 &
[[email protected] ~]# jobs
[1] Running firefox &
[2]- Has stopped vim 1.sh
[3]+ Has stopped vim 2.sh
[4] Running dd if=/dev/zero of=/bigfile bs=1M count=4000 &
[[email protected] ~]# Recorded 4000+0 Read in of
Recorded 4000+0 Write
4194304000 bytes (4.2 GB, 3.9 GiB) copied, 178.781 s, 23.5 MB/s
[4] Completed dd if=/dev/zero of=/bigfile bs=1M count=4000
[[email protected] ~]# fg %2 # The background task is transferred to the foreground
vim 1.sh
[[email protected] ~]# jobs
[1]- Running firefox &
[3]+ Has stopped vim 2.sh
[[email protected] ~]# kill -9 %3 # Kill the background task
[3]+ Have killed vim 2.sh
[[email protected] ~]# jobs
[1]+ Running firefox &
nice value
PRI： The priority given by the system cannot be manually modified
NI：nice value , It can be considered that the scope of change -20～19
1 root You can change your own or others' programs Nice Range of values -20～19
2、 General users can only adjust their own Nice Range of values 0-19
3、 General users can only nice The higher the value, the higher
Finally, the priority formula ：PRI（ new ）=PRI（ used ）+nice value
The lower the priority, the first to execute
adjustment nice Value instruction ：
System settings nice Value to use ：nice command
There are already assigned nice Value to use ：renice command
grammar ：
nice -n Numbers command
-n A number range can be followed -20～19
nice -n -5 vi & Start a process and set nice value
renice 10 20485 Modify an already running program
Example ：
nice -n -5 vi &
[[email protected] desktop ]# ps -l
F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME CMD
4 S 0 18949 18947 0 80 0 - 1719 - pts/0 00:00:00 bash
4 T 0 20485 18949 0 75 -5 - 1700 - pts/0 00:00:00 vi
0 T 0 20490 18949 0 80 0 - 1700 - pts/0 00:00:00 vi
4 R 0 20491 18949 0 80 0 - 1624 - pts/0 00:00:00 ps
renice 10 Process number , Then the modified value directly becomes the default value 80+10
[[email protected] desktop ]# renice 10 20485
20485: old priority -5, new priority 10
[[email protected] desktop ]# ps -l
F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME CMD
4 S 0 18949 18947 0 80 0 - 1719 - pts/0 00:00:00 bash
4 T 0 20485 18949 0 90 10 - 1700 - pts/0 00:00:00 vi
0 T 0 20490 18949 0 80 0 - 1700 - pts/0 00:00:00 vi
4 R 0 20778 18949 3 80 0 - 1625 - pts/0 00:00:00 ps
