- Install virtualBox and vagrant to automate the process of creating a virtual machine.
- Take a look at vagrant cloud to see what public vagrant boxes are available
- (Optional) Customise Your terminal and vim
- Zoom the font size of your terminal:
Terminal -> Preferences -> Switch to
texttab -> Font - Set terminal colour in ~/.bash_profile:
# Global terminal colours
export CLICOLOR=1
export LSCOLORS=GxFxCxDxBxegedabagacedH
- Show git branch on terminal in ~/.bash_profile:
# Git branch in prompt.
parse_git_branch() {
git branch 2> /dev/null | sed -e '/^[^*]/d' -e 's/* \(.*\)/ (\1)/'
}
export PS1="\u@\h \W\[\033[31m\]\$(parse_git_branch)\[\033[00m\] $ "
- Set vim colour in ~/.vimrc:
syntax on
colorscheme desert
- Display line numbers in vim:
Vim any file
$vim ANY_FILEType :set nu and enter
-
Useful shortcuts
- terminal
- Search command history:
ctrl+R, next:R
- Search command history:
- vim
- Jump to line 14:
:14 - Scroll to top:
ggor:1 - Reach the end of a file:
shift+g - Search:
/or?
- Jump to line 14:
- terminal
- Download a box(OS image) and store it to your local storage , e.g.
$vagrant box add username/os_name- Initialise a vagrant project and import the box into VirtualBox:
$vagrant init username/os_name
$vagrant up-
You can create a box with your very own Vagrantfile, which contains a serial of steps of creating a machine. (Think of it a script defines commands of VirtualBox to create a virtual machine)
-
Once you have the machine created, what you do the most might be to start the machine and SSH into the machine.
- Run the box (with the username test1 for example)
$vagrant up test1- SSH into the machine
$vagrant ssh test1And you will see something like this [vagrant@testbox01 ~]$.
- Put shebang in the first line
#!/bin/bashWhen you write shebang, i.e. "#!", it specifies which interpreter this script is going to use.
- To see what a command actually is, check its type at first.
Example 1, a program
$type -a whoamiAnd you will get whoami is /usr/bin/whoami. It is a program, so we can open its manual by typing man
$man whoamiExample 2, a shell keyword
$type -a ifIt turns out if is a shell keyword. So now you can learn the keyword by typing help
$help ifExample 3, a shell builtin
$type -a testIt turns out:
test is a shell builtin
test is /usr/bin/test
Show the manual by typing help, you can also add | less to get it more readable.
$help test | less- To know what command we are using, for example, if I type
$head, where is thisheadfrom
$which headAnd you will see /usr/bin/head
- Another example is
userdel, if you run
$type -a userdelthen you get
userdel is /usr/sbin/userdel
userdel is /sbin/userdel
userdel is /usr/sbin/userdelYou can use which command to check waht userdel you are using right now.
- Check access permission at first
$ls -lAnd you will see the permission
-rw-r--r-- 1 vagrant vagrant 32 Dec 3 08:19 test1_echo.shIgnore the first -, we have three characters a group, the first rw- is the permission of the owner of the file means the readable, writeable but non-executable. The next set r-- represents the permission of the group of the file, the last r-- represents the permission that everyone else in this system granted.
- Execute the file and you will get Permission denied error
- Grant the permission to execute the file
$chmod 755 test1_echo.shr=4, w=2, x=1 To calculate the number of
chmod, for example, I needrwxr-xr-x, the number will be (4+2+1) (4+1) (4+1) = 755. That's how 755 comes from.
Type ls -l again, and you'll see:
-rwxr-xr-x 1 vagrant vagrant 32 Dec 3 08:19 test1_echo.sh- Execute the script
$./test1_echo.sh.: this directory /: separate the file and
...: parent directory
You can simply run the file by ./test1_echo.sh, or you can even go from parent directory like ../localusers/test1_echo.sh
To see if a command is a shell builtin, type
$type [COMMAND]For example, you can try $help echo, and you will get echo is a shell builtin
You can get documentation of any shell builtin by typing
$help [SHELL_BUILTIN]To show in another page
$help [SHELL_BUILTIN]|lessE.g. $help echo
To print something such as variables on the terminal.
#!/bin/bash
# Print 'Hello world'
echo 'Hello World'
# Assign a value to a variable (NO BLANKS before the value)
NAME='Catherine'
# Print the variable (You MUST use double quotes)
echo "$NAME"
# Print a sentence contained a variable
echo "Hello, there! I'm $NAME."
# Combine variables
LINE1='England is barely big enough to contain her. '
LINE2='She will travel Paris, Italy, the Pyrenees. She was mentioning Russia.'
echo "${LINE1}${LINE2}"
# Reassignment
LINE1='Will she be staying long? '
LINE2='Oh, I doubt it.'
echo "${LINE1}${LINE2}"┌──────┐
│ echo │
├──────┴─────────────────────────────────────────────────────────────────┐
│1. No blanks │
│2. Single quotes for value, double quotes for variable │
└────────────────────────────────────────────────────────────────────────┘
Here are the list of special variables this tutorial mentioned:
${UID}${EUID}${0}${?}${RANDOM}${PATH}
To see more bash variables, type man bash
- UID/EUID
$echo "Your UID is ${UID} and EUID is ${EUID}"- id(uid, gid and group)
$idYou will see uid=1000(vagrant) gid=1000(vagrant) groups=1000(vagrant) in Vagrant machine, and you will see more information on Mac by the way (link).
To print one piece of information e.g. name of gid, type:
$id -g -nThese options can be merged. E.g. $id -gn
They're different when a program is running set-uid. Effective UID is the user you changed to, UID is the original user. (Which means UID IS READ ONLY)
- username
USERNAME=$(id -un) # or `id -un` or $(whoami) or `whoami`
echo "Your username: ${USERNAME}"- Check if it is root
if [[ "${UID}" -eq 0 ]]
then
echo 'You are root'
else
echo 'You are not root'
fi- Root user
Two ways to become a root user:
- Run this code snippet with
sudo, and you will get 'You are root' result. - Or you can switch to root user by typing
suat first.
- Run this code snippet with
When you are not root, your might see [vagrant@testbox01 localusers]$, once you become the root user, it shows [root@testbox01 localusers]#
- Display what user typed on the command line. E.g. in test.sh
#!/bin/bash
echo "You executed this command: ${0}"and you run ./test.sh, you will see ./test.sh, but if you run this file by typing /vagrant/localusers/test.sh, it will print /vagrant/localusers/test.sh
┌───────────┐
│ Variables │
├───────────┴────────────────────────────────────────────────────────────┐
│1. Define a code snippet with $(YOUR_COMMAND) or `YOUR_COMMAND` │
│2. Single quotes for value, double quotes for variable │
└────────────────────────────────────────────────────────────────────────┘
readonly PI='3.14'
echo "${PI}"Rule 1
if [[ xxx -eq xxx ]]
then
# do something
fi;E.g.
Write in one single line and separate each command by ;
$if [[ 'a' -eq 'a' ]]; then echo 'same'; fi;To be readable, you can truncate commands by pressing shift + enter. No ; anymore.
$if [[ 'a' -eq 'a' ]]
> then echo 'same'
>fi
-egor=: equal
-neor!=: not equal-lt: less than, which comes with a number.-gt: great than, which comes with a number.
┌────┐
│ If │
├────┴───────────────────────────────────────────────────────────────────┐
│ if [[ condition1 ]] │
│ then │
│ // do something │
│ elif [[ condition2 ]] │
│ then │
│ // do something │
│ else │
│ // do something │
│ fi │
└────────────────────────────────────────────────────────────────────────┘
You may have code snippet like this:
#!/bin/bash
if [[ "${1}" = '-attack' ]] || [[ "${1}" = '--a' ]]
then
echo 'slashing...'
elif [[ "${1}" = '-move' ]] || [[ "${1}" = '--m' ]]
then
echo 'teleporting...'
elif [[ "${1}" = '-bombardment' ]] || [[ "${1}" = '--b' ]]
then
echo 'letting out an arcane torrent...'
elif [[ "${1}" = '-provoke' ]] || [[ "${1}" = '--p' ]] || [[ "${1}" = '-shout' ]] || [[ "${1}" = '--s' ]]
then
echo 'transforming your skin to diamond...'
else
echo 'No idea' >&2 # STDERR
exit 1 # fail
fiAnd you type:
$./test.sh --p
# transforming your skin to diamond...
$./test.sh -sneak
# no idea
$ echo $?
# 1 (from exit status)You can update your to case version
#!/bin/bash
case "${1}" in
-attack|--a)
echo 'slashing...'
;;
-move|--m)
echo 'teleporting...'
;;
-bombardment|--b)
echo 'letting out an arcane torrent...'
;;
-provoke|--p|-shout|--s)
echo 'transforming your skin to diamond...'
;;
*)
echo 'No idea' >&2 # STDERR
exit 1 # fail
;;
esacWith OPTIND, you can easily handle invalid arguments.
┌──────┐
│ case │
├──────┴─────────────────────────────────────────────────────────────────┐
│ case condition in │
│ variable1) │
│ // do something │
| ;; │
│ variable2) │
│ // do something │
| ;; │
│ *) │
│ // do something │
| ;; │
│ esac │
│ │
│ Argument validation: OPTIND │
└────────────────────────────────────────────────────────────────────────┘
To check if your command works properly, you don't need to see the STDOUT or STDERR, you should check the exit status.
# 0: success; non-zero: error
$echo ${?}- 0: success
- Not 0: fail
- Check existed exit status, take
useraddfor example
$man useradd-
Search the keyword "exit" by typing
/exit,nfor next andNfor previous -
You will find
EXIT VALUES
The useradd command exits with the following values:
0 success
1 can't update password file
2 invalid command syntax
...
- Verify the command we've just typed (in shall script)
USERNAME=$(id -un)
# Test if the command worked (check the exit status of the previous command, i.e. 'id -un' in this case
if [[ "${?}" -ne 0 ]]
then
echo "The id command did not work successfully."
exit 2
fi
echo "Username: ${USERNAME}"- Verify the command we've just typed (in bash)
$id -unThen you get "vagrant"
$echo "${?}"And you get 0
Now if I run a false command like
$id -abcYou get the error:
id: invalid option -- 'b'
Try 'id --help' for more information.
Get the exit status, and this time, it will be 1
$echo "${?}"┌─────────────┐
│ Exit status │
├─────────────┴──────────────────────────────────────────────────────────┐
│ > status code │
│ 0: success │
│ others: fail │
│ │
│ > At the end of your shell script, add: │
│ exit 0 │
└────────────────────────────────────────────────────────────────────────┘
A file descriptor is simply a number that represents a open file.
By default, every new process starts with 3 open file descriptors:
- FD 0: STDIN
- FD 1: STDOUT
- FD 2: STDERR
By default, a STDIN comes from your keyboard, and STDOUT and STDERR are displayed on the screen. None of them are files. In fact, Linux represents practically everything as a file.
Aka. STDIN
- Read the input with comments
$read -p 'type something: ' WORDS
type something: Hello
$echo "${WORDS}"
Hello- Read one line of STDIN
# Create a file at first
FILE="tmp"
echo "hello" > ${FILE}
# Read the first line of the tmp file
read LINE_1 < ${FILE}
echo "Line 1: ${LINE_1}"- Read implicitly or explicitly
# Read files in an implicit way (using default file descriptor (0))
$read x < /etc/centos-release
$echo "${x}"
# Read files in an explicit way (NOTICE, no space between 0 and <)
$read x 0< /etc/centos-release
$echo "${x}"Aka. STDOUT
Formula: STATEMENT1 TYPE> STATEMENT2
Formula: STATEMENT1 TYPE>> STATEMENT2
STATEMENT1: a statement that will through info
TYPE: STDOUT is1>or>; STDERR is2>; Both is&>; Convert STDOUT to STDERR by using1>&2, the converse is2>&1
>vs.>>:>is to overwrite whereas>>is to append to the next line.
STATEMENT2: Somewhere to keep outputs orcatstatement to show on screen.
- Write the output to a file.
FILE="tmp"
head -n3 /etc/passwd > ${FILE}You will get tmp file filled in
root:x:0:0:root:/root:/bin/bash
bin:x:1:1:bin:/bin:/sbin/nologin
daemon:x:2:2:daemon:/sbin:/sbin/nologin
- Append string to next line.
echo $(date | sha256sum | head -c10) >> ${FILE}
echo "end" >> ${FILE}The tmp file will be written
root:x:0:0:root:/root:/bin/bash
bin:x:1:1:bin:/bin:/sbin/nologin
daemon:x:2:2:daemon:/sbin:/sbin/nologin
84ba4cc4cf
end
- Write implicitly or explicitly
# Write files in an implicit way (using default file descriptor (1))
$echo "${UID}" > uid
$cat uid
# Read files in an explicit way (NOTICE, no space between 1 and >)
$echo "${UID}" 1> uid
$cat uidTo check if your command works properly, you don't need to see the STDOUT or STDERR, you should check the exit status.
# 0: success; non-zero: error
$echo ${?}- Prerequisites
Here is a head command, it prints the first line of designated files
$head -n1 /etc/passwd /etc/hostsIt will print:
==> /etc/passwd <==
root:x:0:0:root:/root:/bin/bash
==> /etc/hosts <==
127.0.0.1 testbox01 testbox01
Now, to make this command malfunction, we add an fake file.
$head -n1 /etc/passwd /etc/hosts fakefileThe first two files still work properly, but it shows an error message underneath.
==> /etc/passwd <==
root:x:0:0:root:/root:/bin/bash
==> /etc/hosts <==
127.0.0.1 testbox01 testbox01
head: cannot open ‘fakefile’ for reading: No such file or directory
If we write the outputs into a file, this error message will not be written.
$head -n1 /etc/passwd /etc/hosts fakefile > head.out
$cat head.outIt will print:
==> /etc/passwd <==
root:x:0:0:root:/root:/bin/bash
==> /etc/hosts <==
127.0.0.1 testbox01 testbox01
- STDERR
But if we specify its file descriptor to 2 (STDERR)
$head -n1 /etc/passwd /etc/hosts fakefile 2> head.err
$cat head.errIt will print:
head: cannot open ‘fakefile’ for reading: No such file or directory
We can merge make this operation a bit fancy. I.e. using STDOUT and STDERR in one command:
$head -n1 /etc/passwd /etc/hosts fakefile > head.out 2> head.errOr you can append error messages by using 2>>
Another common feature is to merge STDOUT and STDERR together:
$head -n1 /etc/passwd /etc/hosts fakefile > head.both 2>&1which is equivalent to the shorter statement below:
$head -n1 /etc/passwd /etc/hosts fakefile &> head.bothRun cat -n to print it with number of lines.
$cat -n head.bothIt will print:
1 ==> /etc/passwd <==
2 root:x:0:0:root:/root:/bin/bash
3
4 ==> /etc/hosts <==
5 127.0.0.1 testbox01 testbox01
6 head: cannot open ‘fakefile’ for reading: No such file or directory
In pine, STDERR cannot be passed directly, you have convert it to STDOUT at first. Both STDOUT and STDERR go through the pine. E.g.
$head -n1 /etc/passwd /etc/hosts fakefile 2>&1 | catIt is equivalent to another statement
$head -n1 /etc/passwd /etc/hosts fakefile |& catFormula: STATEMENT1 TYPE> STATEMENT2
Formula: STATEMENT1 TYPE>> STATEMENT2
STATEMENT1: a statement that will through info
TYPE: STDOUT is1>or>; STDERR is2>; Both is&>; Convert STDOUT to STDERR by using1>&2, the converse is2>&1
>vs.>>:>is to overwrite whereas>>is to append to the next line.
STATEMENT2: Somewhere to keep outputs orcatstatement to show on screen.
┌─────────────────────────────────┐
│ Standard Input / Output / Error │
├─────────────────────────────────┴──────────────────────────────────────┐
│ > 0: STDIN │
│ 1: STDOUT │
│ 2: STDERR │
│ │
│ > STDIN: │
│ $read x 0< /etc/centos-release │
│ = $read x < /etc/centos-release │
├────────────────────────────────────────────────────────────────────────┤
│ > Formula: `STATEMENT1 TYPE> STATEMENT2` │
│ > STATEMENT1: a statement that will through info │
│ > TYPE: STDOUT is `1>` or `>`; STDERR is `2>`; Both is `&>`; │
│ Convert STDOUT to STDERR by using `1>&2`, the converse is `2>&1` │
│ > `>` vs. `>>`: `>` is to overwrite whereas `>>` is to append to the │
│ next line. │
│ > STATEMENT2: Somewhere to keep outputs or `cat` statement to show │
│ on screen. │
│ │
│ > STDOUT: │
│ > No STDERR │
│ $head -n1 /etc/passwd /etc/hosts fakefile > head.both │
│ = $head -n1 /etc/passwd /etc/hosts fakefile 1> head.both │
│ > Combine STDERR │
│ $head -n1 /etc/passwd /etc/hosts fakefile &> head.both │
│ > Send STDOUT to STDERR │
│ $echo 'error' 1>&2 | cat -n │
│ = $echo 'error' >&2 | cat -n │
│ > Hide STDOUT │
│ $head -n1 /etc/passwd /etc/hosts fakefile 1> /dev/null │
│ = $head -n1 /etc/passwd /etc/hosts fakefile > /dev/null │
│ │
│ > STDERR: │
│ > STDERR Only │
│ $head -n1 /etc/passwd /etc/hosts fakefile 2> head.err │
│ > Send STDERR to STDOUT (In pine, STDERR cannot be passed directly) │
│ $head -n1 /etc/passwd /etc/hosts fakefile 2>&1 | cat │
│ = $head -n1 /etc/passwd /etc/hosts fakefile |& cat │
│ > Hide STDERR │
│ $head -n1 /etc/passwd /etc/hosts fakefile 2> /dev/null │
│ > Hide STDOUT and STDERR │
│ $head -n1 /etc/passwd /etc/hosts fakefile &> /dev/null │
└────────────────────────────────────────────────────────────────────────┘
Log something
$logger 'hello from vagrant user'Print all logs
$sudo tail /var/log/messagesYou will see messages below:
Mar 28 21:58:01 testbox01 vagrant: hello from vagrant user
You can tag on your messages by using -t
$logger -t Doris "Dinner's ready"Again, this time you will see
Mar 28 21:59:09 testbox01 Doris: Dinner's readyList all available checksum methods
$ls -l /usr/bin/*sum
# -rwxr-xr-x 1 root root 33152 Aug 20 02:25 /usr/bin/cksum
# -rwxr-xr-x 1 root root 41504 Aug 20 02:25 /usr/bin/md5sum
# -rwxr-xr-x 1 root root 37448 Aug 20 02:25 /usr/bin/sha1sum
# -rwxr-xr-x 1 root root 41608 Aug 20 02:25 /usr/bin/sha224sum
# -rwxr-xr-x 1 root root 41608 Aug 20 02:25 /usr/bin/sha256sum
# -rwxr-xr-x 1 root root 41624 Aug 20 02:25 /usr/bin/sha384sum
# -rwxr-xr-x 1 root root 41624 Aug 20 02:25 /usr/bin/sha512sum
# -rwxr-xr-x 1 root root 37432 Aug 20 02:25 /usr/bin/sum-b, --binary
read in binary mode
-c, --check
read SHA256 sums from the FILEs and check them
--tag create a BSD-style checksum
-t, --text
read in text mode (default)
Note: There is no difference between binary and text mode option on GNU system.
Get the sha256sum of a file
$sha256sum YOUR_FILEGenerate random numbers
echo ${RANDOM}
# 3054To dive into the random method, go to the Password Generator section
- Print the first line of a file
# 1st line
$head -n1 FILE_NAME
# the first two lines
$head -n2 FILE_NAME- Print the first character of a file
# 1st character
$head -c1 FILE_NAME
# the first three characters
$head -c3 FILE_NAME- Print the first 5 characters of inputs
$echo 12345678 | head -c5- Another feature of
headcommand is to read multiple files. E.g.
$head -n1 /etc/passwd /etc/hostsAnd you will get:
==> /etc/passwd <==
root:x:0:0:root:/root:/bin/bash
==> /etc/hosts <==
127.0.0.1 testbox01 testbox01
With |, you can manipulate the input stream.
E.g. Print the first 5 characters of inputs
# You will get 12345
$echo 12345678 | head -c5And you can keep modify the inputs with multiple |
E.g. Print the first 8 characters of the sha256 checksum of the timestamp
$date +%s | sha256sum | head -c8wrap each input line to fit in specified width
-b, --bytes
count bytes rather than columns
-c, --characters
count characters rather than columns
-s, --spaces
break at spaces
-w, --width=WIDTH
use WIDTH columns instead of 80
For example
S='!@#$%^&*()_-+='
echo "${S}" | fold -b1You will get these characters printed on a single line one by one
!
@
#
$
%
^
&
*
(
)
_
-
+
=
A further use case of fold is to generate a random string. The shuf method works on shuffling lines of text, you cannot shuffle a string directly, you must fold it at first.
echo "${S}" | fold -c1 | shuf#
+
=
-
@
(
$
!
*
_
^
&
%
)
And you can print each line of the character by using head
E.g. I have a file in /vagrant/localusers/test_path.sh
basename: Get the file name (removed the directory)
$basename /vagrant/localusers/test_path.shIt returns test_path.sh
dirname: Get the path of a file (except the file itself)
$dirname /vagrant/localusers/test_path.shIt returns /vagrant/localusers
To find a command which is not in your path, you need locate, install and configure.
locate searches an index created by the updatedb command, which is typically scheduled to run once a day. Therefore, locate command does not have up-to-the-minute information.
- Locate
userdel(You may be able to see more results by runninglocatewithsudo)
$locate userdelthen you would get
/usr/sbin/luserdel
/usr/sbin/userdel
/usr/share/bash-completion/completions/userdel
/usr/share/man/de/man8/userdel.8.gz
/usr/share/man/fr/man8/userdel.8.gz
/usr/share/man/it/man8/userdel.8.gz
/usr/share/man/ja/man8/userdel.8.gz
/usr/share/man/man1/luserdel.1.gz
/usr/share/man/man8/userdel.8.gz
/usr/share/man/pl/man8/userdel.8.gz
/usr/share/man/ru/man8/userdel.8.gz
/usr/share/man/sv/man8/userdel.8.gz
/usr/share/man/tr/man8/userdel.8.gz
/usr/share/man/zh_CN/man8/userdel.8.gz
/usr/share/man/zh_TW/man8/userdel.8.gz- Create your own userdel
$touch userdelIf you run locate userdel, you won't see the userdel you placed.
- Scan the whole file directory to update the DB
$sudo updatedb- Locate
userdelagain, and this time, you will get the latest results.
$locate userdel- To narrow down the search results, you could user
| grep xxxto print results with keywordsxxx
$locate userdel | grep binTo see where a command from, you can type which. E.g.
$which headThen you will see /usr/bin/head
- To create my own
headshell script
$vim user/local/bin/headAnd fill in something like
#!/bin/bash
echo 'Hello from my head'- Escalate the privilege
$sudo chmod 755 /usr/local/bin/head- Check if your head command is changed
$which headAnd you will see /usr/local/bin/head. Besides, you can list all matches by typing $which -a head.
- Run the
headcommand
$head # Hello from my head- You can still run the previous head by assigning the full path. E.g.
$/usr/bin/head -n1 /etc/passwd- Remove the
headand stop using this user-defined shell script.
$sudo rm /usr/local/bin/head- Run
headto see if it works properly. If not, run
# This command is to forget all remembered locations
$hash -rSee how many arguments user inputs
NUMBERS_OF_PARAMS="${#}"
echo "You supplied ${NUMBERS_OF_PARAMS} argument(s) on the command line"The simplest way to get input arguments is "${1}".
E.g. in test.sh, you type $./test.sh A B C D E
echo "Parameter 1: ${1}"
echo "Parameter 2: ${2}"
echo "Parameter 3: ${3}"You will get
Parameter 1: A
Parameter 2: B
Parameter 3: C
You will lose rest of the arguments. In order to get all arguments, you need a for-loop or while-loop.
Three ways to handle user inputs:
read: Pause the process and ask for inputs- input arguments: User types arguments while executing the shell script
- input options: User types arguments while executing the shell script
The difference between input arguments and input options is the syntax user inputs. E.g.
A command with arguments will be like ./test.sh arg1 arg2
On the other hands, a comment with options will be like ./test.sh -v -i
Get all arguments
for ARGUMENT in "${@}"
do
echo "${ARGUMENT}"
doneIf you want to gather all input characters including whitespace as one argument, you can do
for ARGUMENT in "${*}"
do
echo "${ARGUMENT}"
donewhile + shift
- Create a script test.sh
while [[ "${#}" -gt 0 ]]
do
echo "Number of parameters: ${#}"
echo "Param 1: ${1}"
echo "Param 2: ${2}"
echo "Param 3: ${3}"
echo
shift
done
-gt: great then
- Run
./test.sh A B C, and then you will get
Number of parameters: 3
Param 1: A
Param 2: B
Param 3: C
Number of parameters: 2
Param 1: B
Param 2: C
Param 3:
Number of parameters: 1
Param 1: C
Param 2:
Param 3:
┌─────────────┬──────────────┐
│ for-loop │ while-loop │
├─────────────┴──────────────┴───────────────────────────────────────────┐
│ > for-loop # print 1-5 │
│ for e in {1..5} │
│ do │
│ echo "${e}" │
│ done │
│ │
│ > for-loop # print 1, 3, 5, 7, 9 │
│ for e in {1..9..2} │
│ do │
│ echo "${e}" │
│ done │
│ │
│ > for-loop # Loop an array │
│ CARS=('sedan' 'suv' 'hatchback') │
│ for e in ${CARS[*]} │
│ do │
│ echo "${e}" │
│ done │
├────────────────────────────────────────────────────────────────────────┤
│ > while-loop │
│ while [[ CONDITION ]] │
│ do │
│ # do something │
│ done │
└────────────────────────────────────────────────────────────────────────┘
Functions need to be defined before they are used.
- Two ways to define a function without arguments
#!/bin/bash
f1() {
echo 'running f1()'
}
function f2 {
echo 'running f2()'
}
f1
f2- Functions with arguments
f3() {
# The `local` command can only be used inside of a function
local ARGS="${@}"
echo "${ARGS} from f3()"
}
f3 'a' 'b'- Using global variable to control the flow of a function
# This is how shell script define a constant
readonly VERBOSE='true'
log() {
if [[ "${VERBOSE}" = 'true' ]]
then
# The `local` command can only be used inside of a function
local MESSAGES="${@}"
echo "${MESSAGES} from log()"
fi
# To see your log, run: sudo tail /var/log/messages
logger -t my_app "${MESSAGES}"
}
log 'exception: 404 not found'getopts parses command line arguments.
Demo: password_generator.sh
The variable optind is the index of the next element to be processed in argv.
Demo: password_generator.sh
To calculate a number and assign it as a variable, you will need:
$NUM=$(( 6 / 4 ))If you print it, you will get 1, and which is not correct
$echo ${NUM}If you need to calculate, you will need to install another program such as bc
$sudo yum install -y bc$bc -hUpdate a number in another statement.
$NUM='2'
$(( NUM++ ))
$echo ${NUM} # NUM = 3
$NUM=$(( NUM +=5 ))
$echo ${NUM} # NUM = 8$let NUM='2 + 3'
$echo ${NUM} # NUM = 5$let NUM++ # NUM = 6$expr 6 / 4 # 1$NUM=$(expr 1 + 1) # NUM = 2Allow floating calculation in bc, you will need to add -l
$bc -l- bc reads standard input, and to calculate floating numbers, you will need
-loption
$echo '6 / 4' | bc -lAnother way to calculate floating numbers
$awk "BEGIN {print 6/4}"$YOUR_COMMAND | grep KEYWORDSE.g. To locate all password files (including root) placed in system directory
$sudo locate password | grep system$!!Or with sudo
$sudo !!if [[ "${UID}" -eq 0 ]]
then
echo 'You are root'
else
echo 'You are not root'
fi- create a new user
$sudo useradd -c leonardo-da-vinci -m da-vinci-c, --comment COMMENT
Any text string. It is generally a short description of the login, and is currently used as the field for the user's full name.
-m, --create-home
Create the user's home directory if it does not exist. The files and directories contained in the skeleton directory (which can be defined with the -k option) will be copied to the home directory.
By default, if this option is not specified and CREATE_HOME is not enabled, no home directories are created.
The directory where the user's home directory is created must exist and have proper SELinux context and permissions. Otherwise the user's home directory cannot be created or accessed.
- Set the password
Using echo
$sudo echo 123456 | passwd --stdin da-vinciOr using STDIN
# Create a password file
echo "123456" > secret
passwd --stdin da-vinci < secret- Expire a password for an account. The user will be forced to change the password during the next login attempt.
$sudo passwd -e da-vinci- Switch to da-vinci, the user we just created
$su - da-vinci- Logout
$exitThe script to create a user: add_local_user.sh
- Rule of thumb: pwd_generator_guideline.sh
- Demo: password_generator.sh
-
The script allow you to:
- Create local user
- Generate a random password for local user
- Return exit status and log
- Print username(s) and password(s) on console
-
Demo: add_local_users.sh
$./add_local_users.sh Kent David Emma-
The script allow you to:
- Create local user
- Generate a random password for local user
- Return exit status (0: Success; 1: failure + STDERR)
- Keep log to a
log.txtfile - Hide STDOUT of statements, and keep STDERR to the log file
- Print username(s) and password(s) on console
-
Demo: add_local_users_prod.sh
$./add_local_users_prod.sh Kent David Emma- Back up
/etc/passwdfile - Define a function to create a backup for files.
- Log message by using system logger
- Notice, files in
/var/tmp/deleted every 10 days whereas files in/var/deleted every 30 days.
#!/bin/bash
readonly VERBOSE='true'
log() {
if [[ "${VERBOSE}" = 'true' ]]
then
# The `local` command can only be used inside of a function
local MESSAGES="${@}"
echo "${MESSAGES} from log()"
fi
# To see your log, run: sudo tail /var/log/messages
logger -t my_app "${MESSAGES}"
}
backup_file() {
local FILE="${1}"
if [[ -f "${FILE}" ]]
then
local BACKUP="/var/tmp/$(basename ${FILE}).$(date +%F-%N)"
log "Backing up ${FILE} to ${BACKUP}."
# The exit status of the function will be the exit status of the cp -p command
cp -p ${FILE} ${BACKUP}
else
# The file doesn't exist, return a non-zero status
return 1
fi
}
backup_file '/etc/passwd'
if [[ "${?}" -eq '0' ]]
then
log 'File backup succeeded'
else
log 'File backup failed'
exit 1
fi
# It will print on terminal: Backing up /etc/passwd to /var/tmp/passwd.2020-03-29-279478030. from log()
# You can see the log by running: sudo tail /var/tmp/messages
# Your backup will be in /var/tmp/passwd.2020-03-29-279478030The difference between
cpandcp -p, for example:
If you typels -l /etc/passwd, you will see
-rw-r--r-- 1 root root 1184 Feb 23 00:43 /etc/passwd
Copy the passwd file and paste it into /tmp/ folder
cp /etc/passwd /var/tmp/
The date of your passwd file will be when you paste it into.
-rw-r--r-- 1 vagrant vagrant 1184 Apr 2 01:23 passwd
Therefore, to fix this issue, you need to add-pto make sure the date of your copy will be exactly the same.cp -p /etc/passwd /var/tmp/
-rw-r--r-- 1 vagrant vagrant 1184 Feb 23 00:43 passwd