Linux Fundamentals

Introduction

Linux Distributions

LinuxFoundation is an organization that takes care of many Linux projects. Like Linux Foundations, there are many other organizations who take care some of Linux projects. For example, Linux Kernel, Systemd, GNU utils are some of these projects. All these software are available for free, and anybody can use it. So if you are very skilled you can take all this software, put it together, and build your own operating system. And that is exactly the essence of what we call distribution. So the Linux distribution is a collection of software that consists of all this open source software.

One of the major and leading Linux distribution is RedHat. For the simple reason that RedHat is providing enterprise support on Linux. There are two related distributions. 1. CentOS 2. Fedora CentOS is the community enterprise operating system where it doesn’t contain any proprietary information which RedHat inc. has. So it’s a community enterprise OS. Fedora is the development platform, or the community operating system. Another important Linux distribution is Ubuntu. Ubuntu is based on old Linux distrubution Debian. Based on Ubuntu, there is one more distrubution called Linux Mint. There are many other distrubutions. Viz. SUSE & Oracle Linux

Linux File System Hierarchy

Every Linux file system starts with /.

/ is the root directory.

/bin for binaries.

/home for user home directories

/var directory, where you will find log files and more. var contains a wide variety of files that dynamically can be created by the operating system, and one of the most commonly known directories in var is var/log. In var/log, you will find log files that your system is creating.

/usr It is probably one of the most important directory on your system. It’s like Program Files in Windows, and you will find all of the binaries, which are divided in the subdirectory bin and sbin. bin for regular binaries, sbin for system binaries. So sbin are root commands, and bin are regular user commands.

/boot directory: for booting, and often, depending on your Linux configuration, the boot directory will be on a separate partition, because the files that are in the boot directory need to be accessible at all times. Linux kernel, the heart of your operating system is also located here.

/dev It contains device files and these files provide access to hardware. One of the device file is hard disk. Fo ex: sda1 is hard disk, sda1 and sda2 are partitions on your hard disk.

/etc is for configuration files. For ex: os-release is a standard configuration file that should exist on all Linux services, and it describes the current version that you are using. password file is one another important configuration file

/media and /mnt are for mounts, for connecting devices. Mount means that you can connect a storage device to a separate directory

/proc One of the most intriguing directories on your system is proc, because proc is an interface to the Linux kernel

For ex: cpuinfo, providing CPU information that the kernel has directly detected from the hardware in your computer. meminfo, which is providing memory information.

/root is the home directory for the root user.

/run is a new temp directory. Processes that temporarily need to write files do that in the run directory.

/srv is typically empty. It’s meant for services, and you can use it to store document roots for services like a web server or an FTP server.

/sys is for hardware information.

/tmp is for temporary files. This is the old directory where /run is the new directory, where processis can write temporary files.

Using Essential tools

Some basic commands

Getting help with man

Essential File Management Tools

Listing files with ls

ls list files (and directories) in the current diectory

ls -a show hidden files also

ls -R list files in current directory and sub directories

ls -l long listing. show all fields/coloumns including permissions

ls -lt sort by modification time, newest first

ls -ltr sort by modification time, newest last

ls -d or ls -ld list directories themselves, not their contents. Useful, when you have to check permission of a directory

ls -il list inode or index number also

ls -S sort files by size

Using wildcards

ls * list everything recursively

ls a* list everything recursively which starts with ‘a’. Note: if a directory starts with ‘a’, all files in it will be listed.

ls -d a?d* list files which starts with ‘a’ and followed by any sigle character and then char ‘d’ followed by any number of characters. For ex: audit

Note: when ‘-d’ is used, it won’t search recursively

ls -d a[nu]*list files which starts with ‘a’ and followed by ‘n’ or ‘u’ and then any characters. For ex: anaconda, audit

ls -d m[a-e]*list files whcih startes with ‘m’ and second char can be either ‘a’ or ‘b’ or ‘c’ or ‘d’ or ‘e’ followed by any characters. For ex: maillog, messages

Copying files with cp

cp anaconda/* /tmp/test This will copy all 6 files to /tmp/test (but not the anaconda dir)

cp anaconda -R /tmp/test This will copy anaconda directory (and it’s files) to /tmp/test

cp anaconda/ -R /tmp/test This will copy all 6 files to /tmp/test (but not the anaconda dir)

cp anaconda/anaconda.log /tmp/test Here, ‘anaconda.log’ file contents will be copied to ‘test’ file

cp anaconda/anaconda.log /tmp/test/ will now copy the file ‘anaconda.log’ to ‘test’ dir

cp anaconda/anaconda.log /tmp/test/test1.log will copy ‘anaconda.log’ file contents to ‘test1.log’

Working with directories

mkdir /tmp/test will create ‘test’ dir in /tmp

mkdir /tmp/test1/test2 would complain mkdir: cannot create directory '/tmp/test1/test2': No such file or directory

mkdir -p /tmp/test1/test2 will create test1 and test2 dir.

rmdir /tmp/test will remove ‘test’ dir, only if it is empty

rm -rf /tmp/test will remove ‘test’ dir and it’s contents

Using absolute and relative paths

Let’s say, current dir is /temp/test/test1

cd ../test2 will switch the current directory to /temp/test/test2. .. is used to refer one folder up. This is relative path.

cd /temp/test/test2 will switch the current directory to /temp/test/test2. This is absolute path

Moving files with mv

mv command is used to move file/dir (cut and paste) and also rename a file/dir

let’s say, /tmp/test is a directory with some files in it.

mv /tmp/test /temp/some-other-dir/test will move ‘test’ directory and it’s contents to ‘some-other-dir’

mv /tmp/test /temp/test-test will rename ‘test’ dir to ‘test-test’

Removing files with rm

rm /tmp/file1 will remove file1

rm -rf /temp/test will remove test directory and it’s contents

Understanding HardLinks and SymbolicLinks

A link is basically like a shortcut which you can see in other operating systems. But on Linux, it’s a little bit different.

There are 2 type of links. hard link and symbolic link (or soft link).

To understand link, you should know about inode. Every file has an inode, and one inode only. In the inode you find all the administration of a file. if you type ls -l you see a lot of information that is the information that is stored in the inode. From the inode, you go to the blocks. The blocks which is the file information that is physically stored on disc. Now in order to get to an inode, for us human beings, we give a name to a file. And the idea is you have a name somewhere in the directory, and this name points to the inode. You can have more than one name on the same inode. So you have name1 and name2 which both can point to the same inode. That’s what we call a hard link. There is no difference between name1 and name2. So there’s no relation between the original one and the link. Both are directly going to the inode. There is no hierarchical difference. No matter which one you remove, the other one will survive. Because both of them use the inode to go to the blocks. So if you write to the file using name1, then you will see that name2 will see the changes as well.

The symbolic link is a little bit different. Let’s say sym1, a name that points to a name1. Here youthere is a hierarchical difference. Because the symbolic link doesn’t go the inode directly, the symbolic go to a hard link. And if this hard link would be removed then the symbolic links become invalid and don’t work anymore. Hard link has two limitations. 1. There is no cross device. 2. hard links cannot be created on directories. So hard links can only be two files, not two directories.

touch /tmp/name1

ls -li name1

903349 -rw-rw-r--. 1 user1 user1 0 May 16 18:04 name1

Here, we created a new file ‘name1’. ls -li is listing the inode number of the file and also the link counter which is 1, as no link is created yet Now, let’s create a hard link.

ln name1 name2

ls -li name1 name2

903349 -rw-rw-r--. 2 user1 user1 0 May 16 18:04 name1

903349 -rw-rw-r--. 2 user1 user1 0 May 16 18:04 name2

As you can see inode number is same for both the files

Create a symlink

ln -s name1 sym1

ls -li name1 sym1

903349 -rw-rw-r--. 2 user1 user1 0 May 16 18:04 name1

903351 lrwxrwxrwx. 1 user1 user1 5 May 16 18:17 sym1 -> name1

Finding files with find

find <path> -name <file-name>

find <path> -user <username> find all files created by a user

mkdir /root/any; find / -user any -exec cp {} /root/any \;

{} -> contains output of first command when used with exec

\; -> exec should always end with \;

find / -size +100m

find /-size +100M 2>/dev/null

Using advance find options

find / -type -f -size +100M

find /etc -exec grep -l any {} \; -exec cp {} root/any/ \;

find /etc -name ‘*’ -type f | xargs grep “127.0.0.1”

Here, xargs is used along with pipe(

) to get the output line by line

Xargs is similar to -exec

Archiving files with tar

tar -cvf my_archive.tar /home

tar -xvf my_archive.tar it will be extracted to the current directory. Use -C to switch the output path

tar -tvf to show the content

Dd if=/dev/zero of=bigfile1 bs=1M count=1024

Dd -> copy block devices

Here, we are copying block device /dev/zero to generate zeros

of -> output file

Managing file compression

using gzip utility:

gzip <filename>

gunzip <file.gz> To uncompress

time <command> to see how much time a command is taking

using bzip2 utility:

bzip2 <filename>

using zip utility:

zip <filename>

file <filename> command will analayze file metadata and tell what type of file it is

tree command will show the file hierarchy

Working with Text Files

Understanding vi

o open a new line

v visualize. Allows you to make a block (select multiple lines)

d delete a block (cut)

p paste

y copy

dd delete a line

u undo

/<string> search

gg to go to top

<shift> g (or G) go to end

:%s/<string>/<new_string>/ search and replace only 1st occurance

:%s/<string>/<new_string>/g search and replace all occurance

vimtutor to learn more about vi or vim

Browsing text files with more and less

more and less are filters for paging through text one screenful at a time. Both does similar job. But less is more powerful than more. So it is recommanded to use less

more <filename>

enter space bar to scroll down. You cannot scroll up when using more

q to quit

less <filename>

here, you can scroll up and down. It’s more intitutive

Using head and tail to see file start and end

head <file> display only first 10 lines

head -n 5 <file> display first 5 lines

tail <file> display last 10 lines

tail -n 1 <file> display last line

head -n 5 <file> | tail -n 1 Display 5th line

tail -f <file> freshen option. realtime display

Displaying file contents with cat and tac

cat -A <filename> Display non printable characters for ex: new line($), space etc

cat -b <filename> Display line numbers skipping empty lines

cat -n <filename> Display line nos including empty lines

cat -s <filename> Remove repeated empty lines

tac <filename Display lines in opposite order

Working with grep

ps aux | grep ssh

grep any /etc* Any files containing ‘any’ in /etc -> it will return file name & the line containing the search

grep -i <search_string> <file> case insensitive

grep -v to exclude search

for ex: ps aux | grep ssh | grep -v grep -> exclude grep

grep -R Recursive

grep -l Just print the file name not the line nos

grep -A3 Print the line containing searched string + 3 lines after that

grep -B3 Print the line containing searched string + 3 lines before that

Understanding regular expression

^ begining of the line $ end of the line \< begining of a word \> end of a word \A start of a file \Z end of a file {n} exact n times {n,} minimal n times {,n} n times max {n,o} between n and o times * zero or more times + one or more times ? zero or one time

Using regular expression with grep

egrep extended grep

some of the search function doesn’t work with grep for which you can use egrep

for ex: man -k user | grep 1|8

here, | should work as or operator, but grep doesn’t understand it. for that use can use egrep

man -k user | egrep 1|8

similarly, egrep ‘ab{2}c’ <filename>

egrep ‘(123)’ <filename> search the block of 123

Using common text processing utilities

cut filter output from a text file

sort sort files, often used in pipes

tr translates uppercase to lowercase

awk search for specific pattern

sed powerful stream editor to batch modify text files

cut -d : -f 1 /etc/passwd filter field no 1 (here the delimeter is :)

cut -d : -f 1 /etc/passwd | sort

echo hello | tr [:lower:] [:upper:]

sed -n 5p /etc/passwd printing line no 5

sed -i interactive

sed -i s/how/HOW/g <filename>

sed -e edit

sed -i -e ‘2d’ <filename> Delete line no 2

awk -F : ‘{ print $4 }’ /etc/passwd print 4th coloumn

sort -n for numeric sorting

awk -F : ‘/any/ { print $4 }’ /etc/passwd print 4th coloumn that contains ‘any’

head -n 5 /etc/passwd | tail -n 1 print 5th line

sed -n 5p /etc/passwd print 5th line

ps aux | awk ‘{ print $1}’ print 1st field/coloumn

cd /etc; grep -l ‘^root’ * 2>/dev/null print only the file names (not the searched string) which has ‘root’ in the line beginning

grep ‘^…$' * 2>/dev/null search files with lines containining only 3 characters

grep ‘\<alex\>’ <filename> search for a specific word. search for ‘alex’ not Alexander

Using su

su switch user. run a command with substitute user and group ID

su - switch to root and change the shell to root’s shell (current directory is home directory of root)

su switch to root, but don’t change the shell (i.e. root’s environment variables will not be available in the current shell though you switched to ‘root’ user)

su - user1 switch to user1 and change the shell to user1’s shell

su user1 switch to user1, but don’t change the shell

Using sudo

sudo <command> Execute the command as the super user or another user

For any user to run sudo, user must be member of wheel group (in centos) or admin group in (Ubuntu). Otherwise, you would get error like user is not in the sudoers file. This incident will be reported

To add a user to wheel group, switch to root user and run,

usermod -aG wheel <user>

(you need to logout and login to take this change in to effect or you can run newgrp -)

When multiple commands needs to be run sudo, you can use:

sudo -i This will ask for the current user’s password (not the root password). If user has the sudo previlage, root shell will be initialized. Then you can run commands without sudo.

In Ubuntu, root user doesn’t have password. So, you cannot use su. The only way to switch user & run commands is sudo -i

Creating simple sudo configuration

sudo configuration is stored in /etc/sudoers which you should never edit manually

You can use the command sudo visudo to edit or update this file

For example, this file contains a line,

%wheel ALL=(ALL) ALL

This line says, for wheel group, ALL hosts, ALL commands, ALL users

Similarly, we can set different rule for different user/group in the sudoers file. For ex:

user1 AAL=/bin/passwd

Here, user1 is allowed change his password, nothing else

We can create snapin file under /etc/sudoers.d dir & create you own configuration

For ex: create a file /etc/sudoers.d/ansible and add the following configuration

ansible ALL=(ALL) NOPASSWD: ALL

With this configuration, ‘ansible’ user will be able to run sudo command without entering the password

Connecting to a Server

Using SSH to connect to remote server

chvt Change virtual terminal

In Linux there are 6 terminals available. Alt+F2, Alt+F3, Alt+F4 etc can be used to toggle between different terminals. terminal 1 is a graphical mode. Rest all are text terminals.

For ex: chvt 2

ssh

ssh-keygen

ssh-copy-id

scp

who shows which terminals are used

Working with Bash Shell

Understanding the shell and other core linux components

Bash(Shell) -> SysCalls -> Kernel -> Hardware

Using I/O redirection and piping

sort < <file_name>

ps aux | tee psfile tee command is used to redirect output to a file as well as to console

Working with history

History of all command is written to ~/.bash_history

history -c clear the current history (from in-memory)

histrory -w write the current history from in memory to history file. In memory hostory will be deleted

Ctrl-R for reverse-i-search

!nn to repeat a specific line from history

Using command line completion

yum install bash-completion install bash completion package if not already installed

Using variables

variable_name=value

echo $variable_name

By default, variables are only known to the current shell. To make it available to the sub shells, use export

export variable_name=value

env to list all environement variables

Using other bash features

Ctr-l clear screen

Ctrl-u wipe out current command line

Ctr-a move to begining of a line

Ctrl-e move to end of line

Ctrl-c interrupt current process

Ctrl-d exit

Ctrl-v end of file sign. can be used to close the editor

Alias set command alias (set through /etc/profile file)

for ex: alias help=man with this, can use help command i.s.o man to get the manual pages

alias egrep=‘egrep —color=auto’

alias fgrep=‘fgrep —color=auto’

alias l.=‘ls -d .* —color=auto’

alias ll=‘ls -l —color=auto’

alias ls=‘ls —color=auto’

alias mv=‘mv -i’

alias rm=‘rm -i’

alias which=‘alias | /usr/bin/which —ty-only —read-alias —show

If you set alias or variables (including export) this in a shell, it will not be permanent. To set it permanenently you can use bash startup files.

Working with Bash startup files

/etc/environment contains list of variables. this file is processed while starting bash. This is universal.

/etc/profile this is for all users (not for processes).Execute while users login

/etc/profile.d snappin dir contains additional configuration

~/.bash_profile user specific version

~/.bash_logout processed when user logsout. can be be used for cleanup tasks

/etc/bashrc processed everytime a subshell is started

~/.bashrc user specific one for sub shell

You can set aliases or export variables in the above confign files

Aliases are available in subshell also, but not the variables. Use ‘export’ to make variables available everywhere

/etc/Profile.d is a directory & you can create .sh files & set your variables or aliases.

User and Group Management

Understanding file ownership

cd /tmp

touch my_file

ls -l my_file

-rw-rw-r--. 1 user1 user1 0 May 16 21:14 my_file

As you can see, file is owned by ‘user1’ and the group owner is ‘user1’. i.e The user who created became owner of the file and the primary group of that user will become group owner of that file.

A default file permission is set. Typically, for files it is 664 (read/write for user & group. For others read) and for directory it is 775

Creating users with useradd

useradd or adduser Creates a user

for ex: useradd -C “my user” -G wheel -s /bin/passwd user1

here -s is setting the default shell to /bin/passwd, -G is adding the user to wheel group & hence giving sudo access

As every time user ‘user1’ tries to login, it will prompt for changing the password. we can only do that with this user

Note: we can set password while creating user but we need to pass encrypted password only. so it is useless

adduser in Ubuntu, will create a user & ask for password. It is interactive

Creating groups with groupadd

groupadd <groupname> create group name

Managing users and groups properties

w will list who logged in to the host

who is shorter version of w

getent is used to get information from administrative databases

getent passwd user1 here administrative database is passwd and it is fetching info about a user called ‘user1’

/etc/shadow is the main configuration file for user accounts

usermod to modifiy user account

for ex: useradd -aG wheel user1 will add user1 to wheel group

Here, a -> is to append to existing set of groups

id : to list the user & group details

groupmod to modify group properties

userdel <username>

groupdel <groupname>

Configuring defaults for new users

useradd -D will list the defaults for a user

/etc/login.defs is used as the default configuration file for users. For example, password age etc.

/etc/default/useradd this is another configuration file

/etc/skell is a directory with files which will be copied to user’s home directory when a new user is created. As a admin we can create files here which needs to be copied to every user’s home

Managing password properties

passwd -l lock the account

password -u unlock the account

password -e expire the password

password -x to set maximum password lifetime

Password -n to set minimum password lifettime

password -S will show password status for a user like what is the age, locked or unlocked etc

echo <password> | passwd —stdin <username>

history -c to clean the history

chage <username> to change the age of a user. it is interactive

Understanding users and group configuration files

/etc/group contains group information.

/etc/passwd traditionally used to contain user & password info. Since it is readable to everyone, password info is now available in which /etc/shadow to edit /etc/passwd file(to add new user), you can use vipw command

Managing current sessions

loginctl allows current session management (alternate to w & who). But it can do many more stuffs.

loginctl list-sessions different sessions that are currently opened

loginctl show-session <id>

loginctl show-user <username>

loginctl terminate-session <session-id>

loginctl session-status <session-id>

Permissions Management

Understanding and managing basic Linux permissions

chown change user owner of a file

For ex:

chown user1 file1

chown user1:user1 file1 here group ownership is also set

chgrp change group owner of a file

chgrp group1 file1

chmod change permission of a file

File permission is represented in 3 or 4 bit number. For example: 775

Here, 7 means read, write & execute, 5 means read & execute.

read - 4, write - 2, execute - 1

chmod 770 file1

chmod g-w file1 This is relative methos of setting permission. Here, g-w means, takeout write permssion from group owner

In directories, these permission settings tranlates to,

read - 4 - list files

write - 2 - create/delete file

execute - 1 - change directory (cd)

Understanding and managing advance Linux permissions

When chmod is used with 4 digits, 1st digit is for suid+guild+sticky bit

suid-4, guid-2, sticky bit -1

suid when suid is set, file will be run as owner of the file irrespective of who runs it

to set suid, you can use chmod u+s <filename>

find / -perm /4000 find all files with suid(set user id) set.

for example, /usr/bin/passwd file is set with 4000 (rwsr,r-x,r-x) with user as root & group as root. When a normal user runs passwd command, it will run as root.

sgid: This is useful in shared user environement. when set on directory, any new files created by a user will have group owner same as the group owner of the directory. By default, group owner of a file will be the primary group of the user who created the file (typically primary group will be same as user name). In shared environment, it will be a problem because other user will not be able to write/have required permissions.

chmod g+s <directory_name> to set sgid

sticky bit: when this is set, only the directory owner or the file owner can delete the files even though users have file removal permission.

chmod +t <directory_name> to set stick bit

directory permission would look like: drwxrws—T

here T represents that, sticky bit is set

s in the middle represents sgid is set

if s is set at the beginning, it says suid is set

Managing umask

Umask is used to set the default permission for both files and directories. It’s a common setting for both files and directories. umask is a shell setting that defines a mask that will be subtracted from the default permission.

default permission on directories are 777

default permission on files are 666

for ex: if you run, umask 022 Any new files you create will have permission as 644

if you run umask 027 Any new directories you create will have as 750

Umask-> if you simply run ‘umask’ it will list the current umask value

It will be stored/set in /etc/profile & bashrc files

Note: typically all system users will have userid below 199. Ordinary user’s user id will be greater than 199.

Storage Management Essentials

Understanding Linux storage solutions

lsblk to list all block devices & partitions. i.e. disks

fdisk <diskname> to create partition. for ex: fdisk /dev/sda

Sda1, sda2 will be like partitions

(You need to use gdsik i.s.o fdisk on Ubuntu which uses new partiioning technique GPT whereas Centos 7 & older version are still using the old techinque MBR)

Understanding MBR and GBT partitions

Creating file systems

After creating partition you need to create file system explicitly in Linux. mkfs utility is used for this.

file system types: .xfs (default file system in Red-hat), .ext4(default in Ubuntu), vfat (compatible for Windows, Linux & Mac), btrfs, rtfs(for Windows support. not really recommanded in Linux)

For ex: mkfs.xfs /dev/sdb1

Mounting file systems

After file system is created, you need to mount it

mount /dev/sdb1 /mnt

simply running mount will list all the mounted devices

umount /mnt Unmount the mount

lsof <path> will show you opened files/processes

findmnt to list the mounts, where it is mounted

!mo will list the commands which starts with ‘mo’ from history

Managing Networking

Understanding IPV4 basics

Understanding IPv6 basics

Applying runtime network configuration

Understanding network device naming

Managing hostnames

/etc/resolv.conf will contain dns server details

dhclient to reach dhcp server and re-obtain the ip

biosdevname reveals info about network device location

hostname -I shows the IP addresses assigned currently to the current host

/etc/hostname will have host name

hostnamectl status about the host (similar to uname -a)

hostnamectl set-hostname set hostname

if no DNS is configured, you can use local alternative, /etc/hosts

here you can have ip & host mapping

/etc/nsswitch.conf will define search order (for ex: local dns 1st & then the dns. local passwd first & then the database)

ping -f -s <no-of-bytes> <ip>

Managing hostname resolutions

Using common network tools

netstat -tulpen everything listnening on the current host

ss (socket stat) is the new alternate to netstat for ex: ss -tuna

nmap (not installed by default. yum install nmap) network scanning utility. will show what ports are available no matter in which machine. It’s a dangerous tool to use if you are not the expert.

dig to check/troubleshooting dns

ip a

ip a a dev tho 1.2.3.4/8

ip route show

uname

ss -tuna will show the port details. but doesn’t show which service is using that.

(it will also show established connection. like any ssh connection established by someone)

cat /etc/services will show service & it’s port details

for ex: gerp 22 /etc/services will show service name for the port 22

Managing Time

Understanding & managing Linux time

hardware time is set in bios

system time

NTP (Network time protocol): synchronize system time from internet

Understanding NTP protocol

Configuring time synchronisation

date will show date & time

date -s set time

hwclock to synchronise hardware time with system time

imedatectl

timedatectl set-time

timedatectl set-timezone

timedatectl status

stratum

insane clock

iBurst

chronyc source wil show ntp servers you are connected with

ntpdate set date from ntp servers

ntpq to query ntp

Working with Systemd

Understanding systemd

systemd is the manager of everything in Linux. It’s the first thing that is started after starting linux kernel. To Boot linux system, first bootlooder will kick off & then bootloader will load the KERNEL & then kernel will start systemd.

systemd is responsible for starting all the services & generating login prompt.

It also starts processes.

It also manages mounts, timer, paths & much more

Items that are managed by systemd are called units.

Default units are in /usr/lib/systemd/system, custom uints are in /etc/systemd

Managing systemd services

systemctl -t help shows unit types (for ex: service, socket, busname,mount,etc)

systemctl list-unit-files shows all installed units

systemctl list-units lists active units

systemctl enable name.service enables but doesn’t start the service

systemctl start name.service starts a service

systemctl disable name.service disable a service

systemctl stop name.service stop a service

systemctl status name.service shows the status

Modifying service configuration

Note: Earlier /etc/fstab was used to manage all mounts. Now systemctl can handle this.

Below listed commands are the examples for configuring services:

yum install vsftdbinstall ftp server

systemctl status vsftdb

systemctl enable vsftdb (to automatically start after reboot)

systemctl start vsftdb

systemctl cat name.service reads current unit configuration

systemctl show shows all available configuration parameters

systemctl edit name.service allows you to edit service configuration

After modifying service configuration, use systemctl daemon-reload

systemctl restart name.service to restart the service

systemctl cat vsftdbd.service this will display the configuration file location. for ex: /usr/lib/systemd/system/vsftdb.service and other details

systemctl show vsfdtb.service

systemctl edit vsftdb.service it will create/edit /etc/systemd/system/vsftdb.service.d/override.conf to overide the original configuration in /usr/lib/systemd/system/vsftdb.service

Understanding targets

A Target is group of services.

Some targets are isolatable, which means you can use them as a state your system should be in.

There are 4 targets available by default

1. Emergency.target
2. rescue.target
3. Multi-user.target
4. Graphical.target

systemctl list-dependencies name.target to see the contents and dependencies of a systemd target

systemctl start name.target

systemctl isolate name.target

systemctl list-dependencies name.target

systemctl get-default

systemctl set-default name.target

systemctl list-units | grep target

Process Management

Understanding Linux processes and jobs

bg (to send jobs to background)

fg (fore ground)

<command> & to immediately send the job to background

Managing interactive shell jobs

Monitoring processes with top

top -u <user>

if you press F, you can see what fields are avaialble in top interface

you can use < and > to sort by different field

z to use different color

W will write current display settings in ~/.toprc

Changing top display properties

Monitoring process properties with ps

ps aux | less

ps aux | grep ssh

ps -ef it will show PPID also

ps fax it will show forest display. i.e. relation b/w processes

ps aux —sort pmem to sort by memory usage

Changing process priority

nice

renice You can use nice & renice from top interaface also

renice -n 5 <pid> here we are setting the priority to 5 for a running process

nice -n -20 <command> & this is for a new process

Managing process with kill

Signal: is a command which is sent to a process.

for ex: SIGHUP, SIGTERM, SIGKILL

kill <pid>

kill -9 <pid> here 9 is the signal no.

When signal number 9 is passed, it will not wait for any opened files to close before killing the process.

killall <process-name> to kill many processes that have a same name

Managing Software

Installing software from source packages

yumdownloader gedit will download gedit rpm package

rpm -iv `

This is the old way. Here we need to download required dependencies manually before installing the intended package

Software manager were developed to fix the dependency problem. They do so by working with repositories.

Common software managers are yum & apt

YUM:

yum uses repositories that are in /etc/yum.repos.d

yum install

yum search doesn’t search deep enough. For that, you can use yum provides

yum remove

yum groups list

yum groups install

yum provides

yum history shows the history. allows you to undo changes

yum list installed list all installed packages

yum update update the entire system

yum repolist shows available repos

GPG key is used to verify package. This will be downloaded along with the package

Understanding software packages

Managing libraries

Understanding repositories

Managing packages with yum

Examples:

yum search nmap list all packages available related to nmap

yum install nmap-frontend

yum list installed nmap will list installed nmap related packages

yum remove nmap

yum search semanage

yum provides semanage

yum groups install “Compute Node”

yum history will show the history with each step

yum history undo <step-name> to undo specific step name. for ex: yum history undo 6

Managing packages with apt

apt-get & apt-cache are old utilities

apt repositories are mentioned in /etc/apt/sources.list

apt search nmap

apt install nmap

apt remove nmap

Using rpm

RPM(redhat package manager); should not be used anymore

dnf Similar to yum, used in fedora. This will be a replacement for yum in future in redhat also.

rpm can be still used to work with RPM database. allows you to package queries

rpm -qf /my/file: will tell which package a file is from

rpm -ql my package queries the database to list package contents

To query package files instead of package contents, add ‘p’ to the options

rpm -qpc my package.rpm lists configuration files in a downloaded package file

rpm -qp —scripts my package.rpm shows scripts that may be present in a package

rpm -qf /usr/bin/passwd

rpm -ql passwd

rpm -qc passwd

yum downloader nc

rpm -qpl <nc-package-downloaded>

yum list all it will list all packages available from al repos

yum list installed will show only the installed packages

Scheduling Tasks

Understanding Linux task scheduling

cron uses crond daemon

crontab -e to edit tasks

Scheduling tasks with cron

systemctl status crond

/etc/crontab

/etc/cron directory

/etc/cron.d is a directory where you can create snapin files

Suppose you want to run a script daily/weekly/monthly etc and you don’t care about the exact time,

then you can place the script in /etc/cron/cron.daily or cron.weekly files

write a message ‘hello’ daily 4pm to syslog

0 16 * * * logger hello

Note: logger command allows to write log directly to syslog

Using systemd timers

A newer alternative to the cron job.

create a .timer unit & run it using systemctl

/usr/lib/systemd/system this will contain some .timer files

timers will look for corresponding service to run.

systemctl status <timer-name>

for ex: systemctl status fstrim.timer

systemctl enable —now fstrim.timer

Using at to schedule tasks

To run a command script only once.

At uses atd daemon

atq to query

atrm to remove job

For example, To create a file 5 min from now, run below commands from terminal

at now +5min touch /tmp/testfile Ctrl d to close at prompt

Reading Log Files

Understanding Linux logging

syslog is the legacy service that takes care of logging. In modern days, it is implemented through rsyslogd

systemd-journald

Working with journalctl

journalctl shows complete journal

journalctl -u <unit> shows info about specific unit (you can use tab completion)

journalctl —dmesg shows kernel messages

combined filters can be used like, journalctl -u crond —since yesterday —until 9:00 -p info

journalctl -u sshd

journalctl tab tab (to list the units by tab completion)

systemctl status sshd gives similar info

Understanding rsyslog

/etc/rsyslog.conf: will contain rules like where speicific log types should be written. for ex: authentication related logs will be written to /var/log/secure