Linux commands

(1) Different Linux commands that achieve the same outcome, (2) Commands that are either equivalent or alike across Windows and Linux systems.

Help for command[edit]

Help for command:

explainshell.com - match command-line arguments to their help text
Linux or macOS : Type command_name --help or man command_name ;
Win : Type command_name /? "Open Source tools Cygwin which provide functionality similar to a Linux distribution on Windows"
Official manpage Ubuntu Manpage

Text file manipulation[edit]

Display the content of text file[edit]

Display the first lines of text file[edit]

head [options] filename (OS: Linux )
vi [options] filename (and press gg) (Linux )
vim [options] filename (and press gg) (Linux ); Open file and press gg for gVim (Win )

Display the first xxx bytes of text file[edit]

On Linux , To display the first bytes characters of text file<ref>cut - get first X characters from the cat command? - Unix & Linux Stack Exchange</ref>

head -c 100 file_name

Display the last few lines of text file[edit]

tail [options] filename (Linux )
tail - cat line X to line Y on a huge file - Unix & Linux Stack Exchange
vi [options] filename (and press G) (Linux )
vim [options] filename (and press G) (Linux ); Open file and press G for gVim (Win )
Tail for Win32 GUI interface (Win ) not support Unicode file for v.4.2.6

Display the entire content of text file[edit]

vi [options] filename (Linux )
cat <filename> (Linux )
type <filename> (Win )

Display the partial content of large text file page by page[edit]

cat <filename> | less (Linux or Cygwin on Win ) Chinese issue<ref>linux - Show special characters in Unix while using 'less' Command - Stack Overflow un-verified</ref>
type <filename> | more (Win ) If the file was encoded in UTF-8, you need to key in chcp 65001 first <ref>(Solved) Window命令視窗中文亂碼改編碼方法</ref>.

Display only certain lines by number[edit]

text processing - With the Linux "cat" command, how do I show only certain lines by number - Unix & Linux Stack Exchange

Show certain lines by line number[edit]

Show certain lines by line number

Newline count (count the number of lines of a file), word count[edit]

wc -l <filename> for Linux <ref>wc (Unix) - Wikipedia, the free encyclopedia</ref> ex: wc -l *.txt to print the newline count of txt files of the current folder
find . -type f -exec cat {} + | wc -l for Linux & macOS to print the newline count of all files on all subdirectories<ref>linux - Use wc on all subdirectories to count the sum of lines - Stack Overflow</ref>
wc -l "$(cygpath -u '<path to file>')" for cygwin on Win ex: wc -l "$(cygpath -u 'c:\Program Files\file.txt')"<ref>Cygwin convert windows path to linux path</ref>

linux - wc -l is NOT counting last of the file if it does not have end of line character - Stack Overflow

Save the console message as text file[edit]

Save the Unix manpage as plain text file

man man | col -b > man.txt Thanks! Save a Unix manpage as plain text

Search text in a file[edit]

grep command for multiple files: grep -ir "string to search" /path/to/directory or grep keyword /path/to/file (Linux )
vim: (1)vim [options] filename (2)press /keyword (Linux )
cat & grep: cat /path/to/file | grep keyword (Linux ) keyword is case sensitive

Search text in a compressed text file

zcat & grep: cat /path/to/compressed_text.gz | grep keyword (Linux )<ref>Unix / Linux: cat .GZ Compressed Text File On Screen</ref>

Search text in files<ref>鳥哥的 Linux 私房菜 -- 學習 bash shell </ref>

Merge multiple plain text files[edit]

Merge multiple text files into one file

File operation[edit]

Find a file by filename[edit]

locate command (Linux ) ex: locate filename to locate the filename quickly. <ref>Linux 快速尋找檔案 - locate - Tsung's Blog</ref>

find command (Linux ) ex: sudo find / -iname filename Find the case-insensitive file name Under the path / (root folder) <ref> Find command: how to ignore case? - Unix and Linux; Find the case-sensitive file name Under the path / (root folder) ex: find / -name filename ref:Tips For Linux - How to find files in Linux using 'find'</ref>

dir command (Win ) ex: dir /s filename <ref>How to find a file in MS-DOS.</ref><ref>/S 顯示指定目錄及所有子目錄中的檔案。 (引用自 dir /? 命令說明)</ref> I have no idea to specify the file path to find the file.

Copy & overwrite file[edit]

Copy and overwrite file without prompt (confirmation)

On Linux cp -f to_be_copy.file to_be_overwrote.file <ref>Linux：檔案複製強制覆寫, cp force overwrite @ 符碼記憶</ref>
On Linux /bin/cp -f to_be_copy.file to_be_overwrote.file
On Win copy /Y C:\source\some.file C:\target\<ref>DOS Command: COPY</ref>

Copy old directory to new directory[edit]

cp -a old_dir new_dir/ or cp -a old_dir/ new_dir/ (Linux )<ref>

from manual:
-a, --archive
              same as -dR --preserve=all
-d     same as --no-dereference --preserve=links
--preserve[=ATTR_LIST]
              preserve the specified attributes (default: mode,ownership,timestamps), if  possible  addi-
              tional attributes: context, links, xattr, all
-R, -r, --recursive
              copy directories recursively

</ref><ref>unix - How to copy with cp to include hidden files and hidden directories and their contents? - Super User</ref>

- If the folder new_dir exists already, system will create the folder old_dir and all files and sub folders under old_dir will copy to the path new_dir/old_dir.
- If the folder new_dir NOT exists, all files and sub folders under old_dir will copy to the folder new_dir.
cp -a old_dir/* new_dir/ (Linux ): All files and sub folders under old_dir will copy to the folder new_dir.
copy /Y old_dir new_dir/ (Win )<ref> /Y 不顯示覆寫現存目的檔案的確認提示。 (引用自 copy /? 命令說明)</ref>
make new directory & copy all old files to new directory
- mkdir new_dir
- cp old_dir/* new_dir
How to move all files in current folder to subfolder? - Ask Ubuntu

Rename (move) files or directory[edit]

mv /path/to/old_folder_name /path/to/new_folder_name (Linux )<ref>Linux Rename Files, Folders or Directories</ref>
mv /path/to/old.file /path/to/new.file (Linux )
mv /path/to/file /path/to/new_folder_name (Linux )
rename old_folder_name new_folder_name (Win )<ref>MS-DOS ren and rename command help</ref>

Delete files or directory[edit]

delete all files or directory

rm -rf /Name_of_directory to delete all files under the 'Name_of_directory (Linux )<ref>[OPTION]
-r remove directories and their contents recursively;
-f ignore nonexistent files, never prompt

(above content cited from linux documentation: rm --help) </ref>

del Name_of_directory (Win )
- After executed del Name_of_directory but the empty directory directory_name will be left.
- Command delete is not available in Win XP
rsync
- quick for many small files (1) create a empty folder ex: mkdit /path/to/source (2) To empty the target folder, keyin rsync -avP --delete /path/to/source/ /path/to/target/ (Linux )
- rsync --delete-before -a -H -v --progress --stats /path/to/empty/folder/ /path/to/target_folder_will_be_deleted/ (Linux )<ref>Linux下使用rsync最快速删除海量文件的方法_果冻的剥壳_新浪博客</ref> un-verified
Using SFTP/FTP to connection to the server and delete selected files. It may costs too much time if there are many files or sub folders!

Delete the files which their file name were matched with the naming rule when the number of files is too many.

find . -name 'naming_rule' -delete or find . -name 'naming_rule' -type f -delete <ref>How to recursively remove .DS_Store files on Mac OS X | Tangential Musings</ref><ref>XYZ的筆記本: Linux 使用 rm 刪除檔案,出現 /bin/rm: Argument list too long</ref><ref>Linux find command</ref>

Download file from remote server[edit]

FTP
wget (with progress bar): wget http://path.to/file
SVN update: svn up http://remote.svn.server/path/ /local/path/ --username=your_account and key in your SVN password <ref>using the www account su - www -c "svn up http://remote.svn.server/path/ /local/path/ --username=your_account --password=xxx" </ref>
more on File transfer methods

File compression and decompression[edit]

compress

pv big.file | gzip > big.file.gz(Linux )<ref>A Unix Utility You Should Know About: Pipe Viewer - good coders code, great reuse</ref> keep the original file
gzip -v big.file(Linux ) NOT keep the original file. File name after compressed: big.file.gz
How do I Compress a Whole Linux or UNIX Directory? for Linux

decompress

gunzip -c big.file.gz > big.file(Linux ) keep the gz file<ref>linux - How do you gunzip a file and keep the .gz file? - Super User</ref>
gzip -d big.file.gz(Linux ) NOT keep the gz file

ref

Linux: tar Extract Files

List directory or files / list directory contents[edit]

ls (Linux ) ls - list directory contents
- List detailed information of files. e.g. ls -lh<ref>鳥哥的 Linux 私房菜 -- Linux 的檔案權限與目錄配置</ref> <ref>[OPTION]
-l use a long listing format
-h, --human-readable; with -l, print sizes in human readable format (e.g., 1K 234M 2G) Quoted from linux man ls</ref>
- List detailed information of files and sort by modification time. ls -lht
- List top N files and sort by modification time. ls -lht | head -N e.g. ls -lht | head -10 First line is the total file size of specified folder.<ref>How to list top 10 files | Unix Linux Forums | UNIX for Dummies Questions & Answers</ref>
ll (Linux ) ll = ls -l
Print the entire path of files and level-unlimited sub-directories in the current directory. On Win using dir/s/b, and on Linux using find /path/to/files/ -print <ref>linux - Making ls output like dir /b /s - Super User</ref>.
create ls command in Win command prompt: echo dir %1 > %systemroot%\system32\ls.bat Thanks, hmjd!<ref>How to create ls in windows command prompt? - Stack Overflow</ref>
List all files and directories including hidden ones in the current directory. On Win is dir /a, and on macOS or Linux is ls -a<ref>How to see hidden files in MS-DOS and the Command Prompt</ref><ref>Bash Shell: Display All Hidden Dot Files In a Directory</ref>.

List only filenames and one file per line

Win dir/b<ref>How to Copy a List of Files in a Windows Folder Into an Excel List</ref>
macOS ls -1<ref>List One Filename Per Line in Linux | Baeldung on Linux</ref>

Verify if a file or a directory exists

On Linux : ls /path/to/file or ls -d /path/to/directory/ e.g. ls -d ~/ to list your home directory<ref>Single linux command to return to home directory</ref>.
On Win :
- dir C:\path\to\file or dir C:\path\to\directory\
- IF EXIST <ref>How to verify if a file exists in a Windows .BAT file? - Stack Overflow</ref><ref>Windows shell string operations (changing backslash to slash) - Stack Overflow</ref>

SET a=C:\path\to\file
REM -- If the directory was separated the \ symbol & the path NOT contains space
IF EXIST %a% (ECHO YES) ELSE (ECHO No)

SET a="C:\path\to\directory contains space\space"
REM -- If the directory was separated the \ symbol & the path contains space
IF EXIST %a% (ECHO YES) ELSE (ECHO No)

SET a=C:/path/to/file
REM -- If the directory was separated the / symbol, the next command replace / with \ symbol.
SET a=%a:/=\%
IF EXIST %a% (ECHO YES) ELSE (ECHO No)

troubleshooting

SET a = C:\path\to\file
REM -- Correct command: SET a=C:\path\to\file
IF EXIST %a% (ECHO YES) ELSE (ECHO No)
REM -- the returned result will not correct!

Print the current directory / where am i[edit]

Linux : pwd print name of working directory. More on pwd.
Win : cd or echo %cd% to print current directory <ref>Windows equivalent to UNIX pwd - Stack Overflow</ref>

Change directory to my home directory[edit]

cd ~/ or cd $HOME (Linux ) will switch to my home directory ex: /Users/ACCOUNT<ref>linux - Where is the $HOME environment variable set? - Super User</ref>
cd %UserProfile% (Win ) will switch to my home directory ex: C:\Users\ACCOUNT more on MS_Windows_Explorer<ref>weka - Where is my home directory located?</ref>

How to list the ten most recent files in the same directory[edit]

How to list the ten most recent files in the same directory

network[edit]

show ip

ifconfig for Linux <ref>鳥哥的 Linux 私房菜 -- Linux 常用網路指令介紹</ref><ref>bash: ifconfig: command not found 無法使用ifconfig指令？先加上 su -l指令</ref> & macOS <ref>mac 要怎麼看自己的ip呢 - Mobile01</ref>
ipconfig/all (Win )

ping You may got no response if the server was disabled the ICMP (Internet Control Message Protocol) response.

ping domain_or_ip (Linux )
ping domain_or_ip -t (Win )

whois: search the information of domain (related: Domain Name Register)

whois domain (Linux or Win : require to install whois for windows)

TRACERT (trace route)

traceroute domain_or_ip (Linux ) & macOS
tracert domain_or_ip -t (Win )

apache operation[edit]

check apache servie is running or not

service --status-all | grep -i httpd (Linux <ref>Red Hat / CentOS: Check / List Running Services</ref><ref>-i means ignore caseHow to use the grep command, by The Linux Information Project (LINFO)</ref>)
netstat -npl | grep httpd for Linux <ref>Linux / UNIX Find Out What Program / Service is Listening on a Specific TCP Port</ref>
netstat -an | find /i "listening" (Win ) find the local open ports & find local address TCP 0.0.0.0:80 is listening<ref>Netstat</ref>
more on connection test...

Where is apache installed?<ref>Enabling and disabling services during start up in GNU/Linux | All about Linux</ref>

cat /etc/rc.d/rc.local if the apache service is already launched

check which mpm (multi processing module) apache is running

httpd -l (Linux & Win ) ex: Linux /usr/local/httpd/bin/httpd -l or Win x:\apache\bin>httpd -l where x:\apache is the installation folder of apache <ref>Check which mpm (multi processing module) apache is running | Binary Tides</ref>

mysql operation[edit]

MySQL commands: 1. Exporting data into MySql sql file & 2. Importing data from MySql sql file

install package[edit]

Linux : ex: install the Pipe Viewer (pv) package
1. How to Enable RPMForge Repository in RHEL/CentOS 7.x/6.x/5.x/4.x
2. (1)yum install pv for RHEL / CentOS / SL / Fedora Linux<ref>Linux / Unix pv Command: Monitor Progress of Data Sent Via a Pipe</ref>; (2) apt-get install pv for Debian / Ubuntu<ref>ivarch.com: Pipe Viewer</ref>;
macOS
1. install Homebrew "package manager for OS X"
2. brew search PACKAGE_NAME ex: brew search wget
3. if the package is available. brew install PACKAGE_NAME ex: brew install wget

System operation[edit]

How to find the location of an executable[edit]

Find the location of an executable e.g. Using where command on unix-like OS

show current time[edit]

date (Linux ) output the system date: Thu Oct 25 15:05:10 CST 2012 <ref>鳥哥的 Linux 私房菜 -- NTP 時間伺服器</ref><ref>DATE (command) - Wikipedia, the free encyclopedia</ref>
TZ=Asia/Taipei date (Linux ) output the time from the Taipei/CST timezone <ref>How can I have `date` output the time from a different timezone? - Unix & Linux Stack Exchange</ref>
date/t (Win ) (parameter: /t will not ask you to change the current date)

specify the location where Linux or MS-DOS looks when using a command[edit]

echo $PATH (Linux )<ref>鳥哥的 Linux 私房菜 -- 檔案與目錄管理 -- 關於執行檔路徑的變數： $PATH</ref>
echo %PATH% (Win ) more on How to setup my system path

reboot the system/server[edit]

reboot (Linux )<ref>reboot - Linux Command - Unix Command</ref>
shutdown -r (Win )<ref>Windows原來也有內建好用的關機軟體(Shutdown.exe) | ㊣軟體玩家</ref>

Shutdown system at a specific date time.

Linux : (1) Enter date to verify the server timezone. (2) Enter sudo shutdown -h 20:02 <ref>How to shutdown Linux at a specific datetime from terminal? - Unix & Linux Stack Exchange</ref>
Win : Schedule Shutdown or Restarts in Windows using the Task Scheduler

OS version[edit]

lsb_release -a on Linux <ref>The Will Will Web | 如何查詢 Linux 的種類與版本 ( Linux Standard Base )</ref>
cat /etc/redhat-release on Linux CentOS<ref>查看操作系统版本的 N 种方式（Windows、CentOS、Ubuntu、Debian） - Rainbowhhy - 博客园</ref>
ver on Win <ref>Find windows OS version from command line</ref>

Show the process list[edit]

Show the process list & kill the high-resource-consumption process<ref>How do I Find Out Linux CPU Utilization? - nixCraft</ref><ref>Show All Running Processes in Linux</ref>

Display the process list for Linux & macOS
- top "display Linux tasks" for Linux or
- top -a or ps aux --sort -rss | more<ref>鳥哥的 Linux 私房菜 -- 第十六章、程序管理與 SELinux 初探</ref> "Sort by memory usage" for Linux
keyin q to leave the process list

Kill the process[edit]

Find which process to be killed. And then keyin the process ID (PID):

kill <PID> or kill -<PID> PID. To kill the process with PID number: 101, enter kill 101. <ref>Kill Process in Linux or Terminate a Process in UNIX / Linux Systems, kill(1): terminate process - Linux man page</ref>
kill -9 <PID>. Force to kill the process with PID number: 101, enter kill -9 101. <ref>鳥哥的 Linux 私房菜 -- 第十六章、程序管理與 SELinux 初探</ref>

Kill the process by process name<ref>Find and kill a process in one line using bash and regex - Stack Overflow</ref>

sudo kill $(ps aux | grep '<NAME>' | awk '{print $2}')

Kill the process by port number<ref>How to kill a process running on particular port in Linux? - Stack Overflow, lsof(8) - Linux man page</ref>

kill -9 $(lsof -ti:3000) 2>/dev/null: Force kill the process using port 3000. The 2>/dev/null suppresses error messages if no process is found. Replace 3000 with your target port number.
Alternative (shorter command): fuser -k 3000/tcp: Kill the process using TCP port 3000 (if fuser is available on your system)

Explanation:

lsof -ti:3000: Find the process ID (PID) using port 3000
- -t: Output only the PID
- -i:3000: Filter by port 3000
2>/dev/null: Redirects error messages (stderr) to /dev/null (discards them) Useful when no process is found on the port to avoid error output
$(lsof -ti:3000): Command substitution<ref>Command Substitution (Bash Reference Manual)</ref>: executes the command inside $() and replaces it with the output
- First runs lsof -ti:3000 to get the PID(s)
- Then substitutes the PID(s) into the kill -9 command
- Example: if port 3000 is used by PID 12345, kill -9 $(lsof -ti:3000) becomes kill -9 12345

Find process running on port[edit]

Find process running on port

Search and extract string from command output[edit]

linux command | awk '/string/' (Linux )<ref>shell script - How to search and extract string from command output? - Unix & Linux Stack Exchange</ref> ex:
- keyin dpkg --get-selections | awk '/tar/' to search the installed package naming tar for Ubuntu.<ref>logical operators in grep</ref>
- keyin dpkg --get-selections | awk '/zip|unzip/' to search the installed package naming zip or unzip for Ubuntu.
linux command | grep string (Linux ) ex: keyin yum list installed | grep tar to search the installed package naming tar for CentOS.<ref>Unix/Linux grep command examples | grep command in Unix and Linux | grep examples | alvinalexander.com</ref>
Windows command | find "string" (Win ) ex: netstat -a | find "3306" (note: enclose string in double quotation marks)<ref>Microsoft Windows XP - Find</ref>

System load average[edit]

Using WQL query to obtain Load capacity of each processor[edit]

on Win :

namespace: ROOT\CIMV2
WQL query: SELECT LoadPercentage FROM Win32_Processor; No need to divide by CPU cores number for Win8
output example: 24 (UInt16, 16 bit Unsigned Integer) unit: %
CPU usage = LoadPercentage; unit: %

Using PHP function to obtain system load average[edit]

on freebsd, Linux & macOS :

PHP function sys_getloadavg() unit: %; Need to divide by CPU cores number

System CPU usage[edit]

Using mpstat command to obtain CPU usage[edit]

Example result of mpstat 2 2 | grep -E '^Average' on Linux

Average:     all    1.31    0.00    0.25    0.00    0.00    0.00    0.00    0.00    0.00   98.43

The output of the mpstat command includes an "Average" line, which provides detailed information about CPU usage, similar to the %Cpu(s) line in the top command but with a slightly different format. Here's an explanation of the columns in the "Average" line:

all: This represents the average values for all CPU cores.

The subsequent columns represent percentages of CPU usage:

The first column (1.31) is the CPU usage by user-space processes, indicating the percentage of time the CPU is occupied by user programs and processes.
The second column (0.00) is the CPU usage by system kernel processes, indicating the percentage of time the CPU is occupied by the operating system kernel.
The third column (0.25) is the CPU usage by processes with a "nice" priority, representing the percentage of CPU time used by processes adjusted with the nice command.
The fourth column (0.00) is the CPU's idle time, indicating the percentage of time the CPU is not executing any tasks.
The fifth column (0.00) is the CPU's time spent waiting for I/O operations to complete, indicating the percentage of time the CPU is waiting for disk or other I/O device responses.
Columns six through eleven (0.00 to 0.00) represent CPU usage in other states, such as the time spent handling hardware or software interrupts and time spent in virtualized environments with CPU steal time.
The last column (98.43) indicates the percentage of time the CPU is idle, representing the time when the CPU is not executing any tasks.

In the above example, the CPU usage is as follows:

User processes have used 1.31% of CPU time.
System kernel processes have used 0.00% of CPU time.
There are no "nice" priority processes using CPU time.
The CPU is idle for 98.43% of the time.
There is no time spent waiting for I/O operations to complete.
There is no time spent handling hardware or software interrupts.
There is no CPU steal time in a virtualized environment.

Using top command to obtain CPU usage[edit]

Command: Using top command e.g. top on Linux or macOS
Command: Using top command e.g. top -l 2 | grep -E '^CPU' on macOS
Command: Using top command e.g. top -n 1 | grep '%Cpu(s)' on Linux

Example result of top -n 1 | grep '%Cpu(s)'

%Cpu(s):  1.5 us,  1.5 sy,  0.0 ni, 96.9 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st

The result provided is the information about CPU usage obtained through the `top` command. It shows the percentages of different CPU states, and each column is explained as follows:

`%Cpu(s)`: This is the header indicating that the following numbers are percentages related to CPU usage.

`us` (User): It represents the CPU usage by user-space processes, which includes the percentage of time the CPU is occupied by user programs and processes. In this case, it's 1.5%, indicating that 1.5% of CPU time is used for running user programs.

`sy` (System): It represents the CPU usage by system kernel processes, indicating the percentage of time the CPU is occupied by the operating system kernel. In this case, it's 1.5%, indicating that 1.5% of CPU time is used for handling operating system kernel tasks.

`ni` (Nice): It represents the CPU usage by processes with a "nice" priority, which includes the percentage of CPU time used by processes adjusted with the `nice` command. In this case, it's 0%, indicating that there are no "nice" processes running.

`id` (Idle): It represents the percentage of time the CPU is in an idle state, indicating the percentage of time the CPU is not executing any tasks. In this case, it's 96.9%, meaning that 96.9% of CPU time is idle.

`wa` (IO Wait): It represents the CPU usage while waiting for I/O operations to complete, indicating the percentage of time the CPU is waiting for disk or other I/O device responses. In this case, it's 0%, indicating that the CPU is not waiting for I/O to complete.

`hi` (Hardware IRQ): It represents the CPU usage caused by hardware interrupts, indicating the percentage of time spent handling hardware device interrupts. In this case, it's 0%, indicating that the CPU is not handling hardware interrupts.

`si` (Software IRQ): It represents the CPU usage caused by software interrupts, indicating the percentage of time spent handling software interrupts. In this case, it's 0%, indicating that the CPU is not handling software interrupts.

`st` (Steal Time): It represents the "steal time" in a virtualized environment, indicating the percentage of time a virtual machine's CPU is taken away when sharing host CPU resources with other virtual machines. In this case, it's 0%, meaning there is no virtualization steal time.

In summary, this output indicates the following CPU usage:

User programs are using 1.5% of CPU time.
System kernel tasks are using 1.5% of CPU time.
There are no "nice" processes running.
96.9% of CPU time is idle.
There is no time spent waiting for I/O operations to complete.
There is no time spent handling hardware or software interrupts.
There is no virtualization steal time.

System memory usage[edit]

Obtaining Memory Usage through a WQL Query[edit]