Beginners' guide

Tip: This is part of a multi-page article for The Beginners' Guide. Click here if you would rather read the guide in its entirety.

Installation

You are now presented with a shell prompt, automatically logged in as root.

Change the language

Tip: These are optional for the majority of users. Useful only if you plan on writing in your own language in any of the configuration files, if you use diacritical marks in the Wi-Fi password, or if you would like to receive system messages (e.g. possible errors) in your own language.

By default, the keyboard layout is set to us. If you have a non-US keyboard layout, run:

# loadkeys layout

...where layout can be fr, uk, dvorak, be-latin1, etc. See here for a comprehensive list.

The font should also be changed, because most languages use more glyphs than the 26 letter English alphabet. Otherwise some foreign characters may show up as white squares or as other symbols. Note that the name is case-sensitive, so please type it exactly as you see it:

# setfont Lat2-Terminus16

By default, the language is set to English (US). If you would like to change the language for the install process (German, in this example), remove the # in front of the locale you want from /etc/locale.gen, along with English (US). Please choose the UTF-8 entry.

Use Ctrl+X to exit, and when prompted to save changes, press Y and Enter to use the same filename.

# nano /etc/locale.gen

en_US.UTF-8 UTF-8
de_DE.UTF-8 UTF-8

# locale-gen
# export LANG=de_DE.UTF-8

Remember, LAlt+LShift activates and deactivates the keymap.

Establish an internet connection

Warning: As of v197, udev no longer assigns network interface names according to the wlanX and ethX naming scheme. If you are coming from a different distribution or are reinstalling Arch and not aware of the new interface naming style, please do not assume that your wireless interface is named wlan0, or that your wired interface is named eth0. You can use the command ip link to discover the names of your interfaces.

The dhcpcd network daemon starts automatically during boot and it will attempt to start a wired connection. Try to ping a server to see if a connection was established. For example, Google's webservers:

# ping -c 3 www.google.com

PING www.l.google.com (74.125.132.105) 56(84) bytes of data.
64 bytes from wb-in-f105.1e100.net (74.125.132.105): icmp_req=1 ttl=50 time=17.0 ms
64 bytes from wb-in-f105.1e100.net (74.125.132.105): icmp_req=2 ttl=50 time=18.2 ms
64 bytes from wb-in-f105.1e100.net (74.125.132.105): icmp_req=3 ttl=50 time=16.6 ms

--- www.l.google.com ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 16.660/17.320/18.254/0.678 ms

If you get a ping: unknown host error, first check if there is an issue with your cable or wireless signal strength. If not, you will need to set up the network manually, as explained below. Once a connection is established move on to Prepare the storage drive.

Wired

Follow this procedure if you need to set up a wired connection via a static IP address.

First, disable the dhcpcd service which was started automatically at boot:

# systemctl stop dhcpcd.service

Identify the name of your Ethernet interface.

# ip link

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: enp2s0f0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether 00:11:25:31:69:20 brd ff:ff:ff:ff:ff:ff
3: wlp3s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DORMANT qlen 1000
    link/ether 01:02:03:04:05:06 brd ff:ff:ff:ff:ff:ff

In this example, the Ethernet interface is enp2s0f0. If you are unsure, your Ethernet interface is likely to start with the letter "e", and unlikely to be "lo" or start with the letter "w". You can also use iwconfig and see which interfaces are not wireless:

# iwconfig

enp2s0f0  no wireless extensions.
wlp3s0    IEEE 802.11bgn  ESSID:"NETGEAR97"
          Mode:Managed  Frequency:2.427 GHz  Access Point: 2C:B0:5D:9C:72:BF
          Bit Rate=65 Mb/s   Tx-Power=16 dBm
          Retry  long limit:7   RTS thr:off   Fragment thr:off
          Power Management:on
          Link Quality=61/70  Signal level=-49 dBm
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:0  Invalid misc:430   Missed beacon:0
lo        no wireless extensions.

In this example, neither enp2s0f0 nor the loopback device have wireless extensions, meaning enp2s0f0 is our Ethernet interface.

You also need to know these settings:

• Static IP address.
• Subnet mask.
• Gateway's IP address.
• Name servers' (DNS) IP addresses.
• Domain name (unless you are on a local LAN, in which case you can make it up).

Activate the connected Ethernet interface (e.g. enp2s0f0):

# ip link set enp2s0f0 up

Add the address:

# ip addr add ip_address/subnetmask dev interface_name

For example:

# ip addr add 192.168.1.2/24 dev enp2s0f0

For more options, run man ip.

Add your gateway like this, substituting your own gateway's IP address:

# ip route add default via ip_address

For example:

# ip route add default via 192.168.1.1

Edit resolv.conf, substituting your name servers' IP addresses and your local domain name:

# nano /etc/resolv.conf

nameserver 61.23.173.5
nameserver 61.95.849.8
search example.com

Note: Currently, you may include a maximum of three nameserver lines. In order to overcome this limitation, you can use a locally caching nameserver like Dnsmasq.

You should now have a working network connection. If you do not, check the detailed Network Configuration page.

Wireless

Follow this procedure if you need wireless connectivity (Wi-Fi) during the installation process.

First, identify the name of your wireless interface.

# iw dev

phy#0
        Interface wlp3s0
                ifindex 3
                wdev 0x1
                addr 00:21:6a:5e:52:bc
                type managed

In this example, wlp3s0 is the available wireless interface. If you are unsure, your wireless interface is likely to start with the letter "w", and unlikely to be "lo" or start with the letter "e".

Note: If you do not see output similar to this, then your wireless driver has not been loaded. If this is the case, you must load the driver yourself. Please see Wireless Setup for more detailed information.

Bring the interface up with:

# ip link set wlp3s0 up

A small percentage of wireless chipsets also require firmware, in addition to a corresponding driver. If you get output like SIOCSIFFLAGS: No such file or directory, this means you will need to manually load the firmware. If unsure, invoke dmesg to query the kernel log for a firmware request from the wireless chipset. For example, if you have an Intel chipset which requires and has requested firmware from the kernel at boot:

# dmesg | grep firmware

firmware: requesting iwlwifi-5000-1.ucode

If there is no output, it may be concluded that the system's wireless chipset does not require firmware.

Warning: Wireless chipset firmware packages (for cards which require them) are pre-installed under /usr/lib/firmware in the live environment (on CD/USB stick) but must be explicitly installed to your actual system to provide wireless functionality after you reboot into it! Package installation is covered later in this guide. Ensure installation of both your wireless module and firmware before rebooting! See Wireless Setup if you are unsure about the requirement of corresponding firmware installation for your particular chipset.

Next, use netctl's wifi-menu to connect to a network:

# wifi-menu wlp3s0

You should now have a working network connection. If you do not, check the detailed Wireless Setup page.

Alternatively, use iw dev wlp3s0 scan | grep SSID to scan for available networks, then connect to a network with:

# wpa_supplicant -B -i wlp3s0 -c <(wpa_passphrase "ssid" "psk")

You need to replace ssid with the name of your network (e.g. "Linksys etc...") and psk with your wireless password, leaving the quotes around the network name and password.

Finally, you have to give your interface an IP address. This can be set manually or using the dhcp:

# dhcpcd wlp3s0

If that does not work, issue the following commands:

# echo 'ctrl_interface=DIR=/run/wpa_supplicant' > /etc/wpa_supplicant.conf
# wpa_passphrase <ssid> <passphrase> >> /etc/wpa_supplicant.conf
# ip link set <interface> up # May not be needed as dhcpcd should bring it up but may be needed for wpa_supplicant.
# wpa_supplicant -B -D nl80211 -c /foobar.conf -i <interface name>
# dhcpcd -A <interface name>

Analog modem, ISDN or PPPoE DSL

For xDSL, dial-up and ISDN connections, see Direct Modem Connection.

Behind a proxy server

If you are behind a proxy server, you will need to export the http_proxy and ftp_proxy environment variables. See Proxy settings for more information.

Prepare the storage drive

Warning: Partitioning can destroy data. You are strongly cautioned and advised to backup any critical data before proceeding.

Choose a partition table type

You have to choose between GUID Partition Table (GPT) and Master Boot Record (MBR). GPT is more modern and recommended for new installations.

• If you want to setup a system which dual boots with windows, then you have to pay special attention to this choice. See Partitioning#Choosing_between_GPT_and_MBR for the gory details.
• It is recommended to always use GPT for UEFI boot, as some UEFI firmwares do not allow UEFI-MBR boot.
• Some BIOS systems may have issues with GPT. See http://mjg59.dreamwidth.org/8035.html and http://rodsbooks.com/gdisk/bios.html for more info and possible workarounds.

Note: If you are installing to a USB flash key, see Installing Arch Linux on a USB key.

Partitioning tool

Absolute beginners are encouraged to use a graphical partitioning tool. GParted is a good example, and is provided as a "live" CD. It is also included on live CDs of most Linux distributions such as Ubuntu and Linux Mint. A drive should first be partitioned and the partitions should be formatted with a file system before rebooting.

Tip: When using Gparted, selecting the option to create a new partition table gives an "msdos" partition table by default. If you are intending to follow the advice to create a GPT partition table then you need to choose "Advanced" and then select "gpt" from the drop-down menu.

While gparted may be easier to use, if you just want to create a few partitions on a new disk you can get the job done quickly by just using one of the fdisk variants which are included on the install medium. There are short usage instructions for both gdisk and fdisk.

Partition scheme

You can decide into how many partitions the disk should be split, and for which directory each partition should be used in the system. The mapping from partitions to directories (frequently called 'mount points') is the Partition scheme. The simplest, and not a bad choice, is to make just one huge / partition. Another popular choice is to have a / and a /home partition.

Additional required partitions:

• If you have a UEFI motherboard, you will need to create an extra UEFI System Partition. It will need to be at least 512 MiB.
• If you have a BIOS motherboard (or plan on booting in BIOS compatibility mode) and you want to setup GRUB on a GPT-partitioned drive, you will need to create an extra BIOS Boot Partition of size 1007 KiB and EF02 type code. Syslinux does not need one.

See Swap for details if you wish to set up a swap partition or swap file. A swap file is easier to resize than a partition and can be created at any point after installation, but cannot be used with a Btrfs filesystem.

Considerations for dualbooting with windows

If you have an existing OS installation, please keep in mind that if you were to just write a completely new partition table to disk then all the data which was previously on disk would be lost.

The recommended way to setup a linux/windows dual booting system is to first install windows, only using part of the disk for its partitions. When you have finished the windows setup, boot into the linux install environment where you can create additional partitions for linux while leaving the existing windows partitions untouched.

Some newer computers come pre-installed with Windows 8 which will be using Secure Boot. Arch Linux currently does not support Secure Boot, but some Windows 8 installations have been seen not to boot if Secure Boot is turned off in the BIOS. In some cases it is necessary to turn off both Secure Boot as well as Fastboot in the BIOS options in order to allow Windows 8 to boot without Secure Boot. However there are potential security risks in turning off Secure Boot for booting up Windows 8. Therefore, it may be a better option to keep the Windows 8 install intact and have an independent hard drive for the Linux install - which can then be partitioned from scratch using a GPT partition table. Once that is done, creating several ext4/FAT32/swap partitions on the second drive may be a better way forward if the computer has two drives available. This is often not easy or possible on a small laptop. Currently, Secure Boot is still not in a fully stable state for reliable operation, even for Linux distributions that support it.

If you have already created your partitions, proceed to Create filesystems.

Otherwise, see the following example.

Example

The Arch Linux install media includes the following partitioning tools: fdisk, gdisk, cfdisk, cgdisk, parted.

Tip: Use the lsblk command to list the hard disks attached to your system, along with the sizes of their existing partitions. This will help you to be confident you are partitioning the right disk.

The example system will contain a 15 GB root partition, and a home partition for the remaining space. Choose either MBR or GPT. Do not choose both!

It should be emphasized that partitioning is a personal choice and that this example is only for illustrative purposes. See Partitioning.

Using cgdisk to create GPT partitions

# cgdisk /dev/sda

Root

• Choose New (or press N) – Enter for the first sector (2048) – type in 15G – Enter for the default hex code (8300) – Enter for a blank partition name.

Home

• Press the down arrow a couple of times to move to the larger free space area.
• Choose New (or press N) – Enter for the first sector – Enter to use the rest of the drive (or you could type in the desired size; for example 30G) – Enter for the default hex code (8300) – Enter for a blank partition name.

Here is what it should look like:

Part. #     Size        Partition Type            Partition Name
----------------------------------------------------------------
            1007.0 KiB  free space
   1        15.0 GiB    Linux filesystem
   2        123.45 GiB  Linux filesystem

Double check and make sure that you are happy with the partition sizes as well as the partition table layout before continuing.

If you would like to start over, you can simply select Quit (or press Q) to exit without saving changes and then restart cgdisk.

If you are satisfied, choose Write (or press Shift+W) to finalize and to write the partition table to the drive. Type yes and choose Quit (or press Q) to exit without making any more changes.

Using fdisk to create MBR partitions

Note: There is also cfdisk, which is similar in UI to cgdisk, but it currently does not automatically align the first partition properly. That is why the classic fdisk tool is used here.

Launch fdisk with:

# fdisk /dev/sda

Create the first partition:

1. Command (m for help): type n and press Enter
2. Partition type: Select (default p): press Enter
3. Partition number (1-4, default 1): press Enter
4. First sector (2048-209715199, default 2048): press Enter
5. Last sector, +sectors or +size{K,M,G} (2048-209715199....., default 209715199): type +15G and press Enter

Then create a second partition:

1. Command (m for help): type n and press Enter
2. Partition type: Select (default p): press Enter
3. Partition number (1-4, default 2): press Enter
4. First sector (31459328-209715199, default 31459328): press Enter
5. Last sector, +sectors or +size{K,M,G} (31459328-209715199....., default 209715199): press Enter

Now preview the new partition table:

• Command (m for help): type p and press Enter

Disk /dev/sda: 107.4 GB, 107374182400 bytes, 209715200 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x5698d902

   Device Boot     Start         End     Blocks   Id  System
/dev/sda1           2048    31459327   15728640   83   Linux
/dev/sda2       31459328   209715199   89127936   83   Linux

Then write the changes to disk:

• Command (m for help): type w and press Enter

If everything went well fdisk will now quit with the following message:

The partition table has been altered!

Calling ioctl() to re-read partition table.
Syncing disks. 

In case this doesn't work because fdisk encountered an error, you can use the q command to exit.

Create filesystems

Simply partitioning is not enough; the partitions also need a filesystem. To format the partitions with an ext4 filesystem:

Warning: Double check and triple check that it is actually /dev/sda1 and /dev/sda2 that you want to format. You can use lsblk to help with this.

# mkfs.ext4 /dev/sda1
# mkfs.ext4 /dev/sda2

If you have made a partition dedicated to swap (code 82), do not forget to format and activate it with:

# mkswap /dev/sdaX
# swapon /dev/sdaX

For UEFI, you should format the ESP partition (usually sda1) with:

# mkfs.vfat -F32 /dev/sda1

Mount the partitions

Each partition is identified with a number suffix. For example, sda1 specifies the first partition of the first drive, while sda designates the entire drive.

To display the current partition layout:

# lsblk /dev/sda

Note: Do not mount more than one partition to the same directory. And pay attention, because the mounting order is important.

First, mount the root partition on /mnt. Following the example above (yours may be different), it would be:

# mount /dev/sda1 /mnt

Then mount the home partition and any other separate partition (/boot, /var, etc), if you have any:

# mkdir /mnt/home
# mount /dev/sda2 /mnt/home

In case you have a UEFI motherboard, mount the UEFI partition:

# mkdir -p /mnt/boot
# mount /dev/sdaX /mnt/boot

Select a mirror

Before installing, you may want to edit the mirrorlist file and place your preferred mirror first. A copy of this file will be installed on your new system by pacstrap as well, so it is worth getting it right.

# nano /etc/pacman.d/mirrorlist

##
## Arch Linux repository mirrorlist
## Sorted by mirror score from mirror status page
## Generated on 2012-MM-DD
##

Server = http://mirror.example.xyz/archlinux/$repo/os/$arch
...

• Alt+6 to copy a Server line.
• PageUp key to scroll up.
• Ctrl+U to paste it at the top of the list.
• Ctrl+X to exit, and when prompted to save changes, press Y and Enter to use the same filename.

If you want, you can make it the only mirror available by getting rid of everything else (using Ctrl+K), but it is usually a good idea to have a few more, in case the first one goes offline.

Tip:

• Use the Mirrorlist Generator to get an updated list for your country. HTTP mirrors are faster than FTP, because of something called keepalive. With FTP, pacman has to send out a signal each time it downloads a package, resulting in a brief pause. For other ways to generate a mirror list, see Sorting mirrors and Reflector.
• Arch Linux MirrorStatus reports various aspects about the mirrors such as network problems with mirrors, data collection problems, the last time mirrors have been synced, etc.

Note:

• Whenever in the future you change your list of mirrors, always remember to force pacman to refresh all package lists with pacman -Syy. This is considered to be good practice and will avoid possible headaches. See Mirrors for more information.
• If you are using an older installation medium, your mirrorlist might be outdated, which might lead to problems when updating Arch Linux (see FS#22510). Therefore it is advised to obtain the latest mirror information as described above.
• Some issues have been reported in the Arch Linux forums regarding network problems that prevent pacman from updating/synchronizing repositories (see [1] and [2]). When installing Arch Linux natively, these issues have been resolved by replacing the default pacman file downloader with an alternative (see Improve Pacman Performance for more details). When installing Arch Linux as a guest OS in VirtualBox, this issue has also been addressed by using "Host interface" instead of "NAT" in the machine properties.

Install the base system

The base system is installed using the pacstrap script.

The -i switch can be omitted if you wish to install every package from the base group without prompting.

# pacstrap -i /mnt base

Note:

• If pacman fails to verify your packages, check the system time with cal. If the system date is invalid (e.g. it shows the year 2010), signing keys will be considered expired (or invalid), signature checks on packages will fail and installation will be interrupted. Make sure to correct the system time, either by doing so manually or with the ntp client, and retry running the pacstrap command. Refer to Time page for more information on correcting system time.
• If pacman complains that error: failed to commit transaction (invalid or corrupted package), run the following command:

# pacman-key --init && pacman-key --populate archlinux

This will give you a basic Arch system. Other packages can be installed later using pacman.

Generate an fstab

Generate an fstab file with the following command. UUIDs will be used because they have certain advantages (see fstab#Identifying filesystems). If you would prefer to use labels instead, replace the -U option with -L.

# genfstab -U -p /mnt >> /mnt/etc/fstab
# nano /mnt/etc/fstab

Warning: The fstab file should always be checked after generating it. If you encounter errors running genfstab or later in the install process, do not run genfstab again; just edit the fstab file.

A few considerations:

• The last field determines the order in which partitions are checked at start up: use 1 for the (non-btrfs) root partition, which should be checked first; 2 for all other partitions you want checked at start up; and 0 means 'do not check' (see fstab#Field definitions).
• All btrfs partitions should have 0 for this field. Normally, you will also want your swap partition to have 0.

Chroot and configure the base system

Next, we chroot into our newly installed system:

# arch-chroot /mnt

Note: Use arch-chroot /mnt /bin/bash to chroot into a bash shell.

At this stage of the installation, you will configure the primary configuration files of your Arch Linux base system. These can either be created if they do not exist, or edited if you wish to change the defaults.

Closely following and understanding these steps is of key importance to ensure a properly configured system.

Locale

Locales are used by glibc and other locale-aware programs or libraries for rendering text, correctly displaying regional monetary values, time and date formats, alphabetic idiosyncrasies, and other locale-specific standards.

There are two files that need editing: locale.gen and locale.conf.

• The locale.gen file is empty by default (everything is commented out) and you need to remove the # in front of the line(s) you want. You may uncomment more lines than just English (US), as long as you choose their UTF-8 encoding:

# nano /etc/locale.gen

en_US.UTF-8 UTF-8
de_DE.UTF-8 UTF-8

# locale-gen

This will run on every glibc upgrade, generating all the locales specified in /etc/locale.gen.

• The locale.conf file does not exist by default. Setting only LANG should be enough. It will act as the default value for all other variables.

# echo LANG=en_US.UTF-8 > /etc/locale.conf
# export LANG=en_US.UTF-8

Note: If you set some other language than English (US) at the beginning of the install, the above commands would be something like:

# echo LANG=de_DE.UTF-8 > /etc/locale.conf
# export LANG=de_DE.UTF-8

To use other locales for other LC_* variables, run locale to see the available options and add them to locale.conf. It is not recommended to set the LC_ALL variable. An advanced example can be found here.

Console font and keymap

If you set a keymap at the beginning of the install process, load it now, as well, because the environment has changed. For example:

# loadkeys de-latin1
# setfont Lat2-Terminus16

To make them available after reboot, edit vconsole.conf:

# nano /etc/vconsole.conf

KEYMAP=de-latin1
FONT=Lat2-Terminus16

• KEYMAP – Please note that this setting is only valid for your TTYs, not any graphical window managers or Xorg.

• FONT – Available alternate console fonts reside in /usr/share/kbd/consolefonts/. The default (blank) is safe, but some foreign characters may show up as white squares or as other symbols. It is recommended that you change it to Lat2-Terminus16, because according to /usr/share/kbd/consolefonts/README.Lat2-Terminus16, it claims to support "about 110 language sets".

• Possible option FONT_MAP – Defines the console map to load at boot. Read man setfont. Removing it or leaving it blank is safe.

See Console fonts and man vconsole.conf for more information.

Time zone

Available time zones and subzones can be found in the /usr/share/zoneinfo/<Zone>/<SubZone> directories.

To view the available <Zone>, check the directory /usr/share/zoneinfo/:

# ls /usr/share/zoneinfo/

Similarly, you can check the contents of directories belonging to a <SubZone>:

# ls /usr/share/zoneinfo/Europe

Create a symbolic link /etc/localtime to your zone file /usr/share/zoneinfo/<Zone>/<SubZone> using this command:

# ln -s /usr/share/zoneinfo/<Zone>/<SubZone> /etc/localtime

Example:

# ln -s /usr/share/zoneinfo/Europe/Minsk /etc/localtime

Hardware clock

Set the hardware clock mode uniformly between your operating systems. Otherwise, they may overwrite the hardware clock and cause time shifts.

You can generate /etc/adjtime automatically by using one of the following commands:

• UTC (recommended)

Note: Using UTC for the hardware clock does not mean that software will display time in UTC.

# hwclock --systohc --utc

To synchronize your "UTC" time over the internet, see NTPd.

• localtime (discouraged; used by default in Windows)

Warning: Using localtime may lead to several known and unfixable bugs. However, there are no plans to drop support for localtime.

# hwclock --systohc --localtime

If you have (or planning on having) a dual boot setup with Windows:

• Recommended: Set both Arch Linux and Windows to use UTC. A quick registry fix is needed. Also, be sure to prevent Windows from synchronizing the time on-line, because the hardware clock will default back to localtime.

• Not recommended: Set Arch Linux to localtime and disable any time-related services, like NTPd . This will let Windows take care of hardware clock corrections and you will need to remember to boot into Windows at least two times a year (in Spring and Autumn) when DST kicks in. So please do not ask on the forums why the clock is one hour behind or ahead if you usually go for days or weeks without booting into Windows.

Kernel modules

Tip: This is just an example, you do not need to set it. All needed modules are automatically loaded by udev, so you will rarely need to add something here. Only add modules that you know are missing.

For kernel modules to load during boot, place a *.conf file in /etc/modules-load.d/, with a name based on the program that uses them.

# nano /etc/modules-load.d/virtio-net.conf

# Load 'virtio-net.ko' at boot.

virtio-net

If there are more modules to load per *.conf, the module names can be separated by newlines. A good example are the VirtualBox Guest Additions.

Empty lines and lines starting with # or ; are ignored.

Hostname

Set the hostname to your liking (e.g. arch):

# echo myhostname > /etc/hostname

Note: There is no need to edit /etc/hosts.

Configure the network

You need to configure the network again, but this time for your newly installed environment. The procedure and prerequisites are very similar to the one described above, except we are going to make it persistent and automatically run at boot.

Note:

• For more in-depth information on network configration, visit Network Configuration and Wireless Setup.
• If you would like to use the old interface naming scheme (ie. eth* and wlan*) you can accomplish this by creating an empty file at /etc/udev/rules.d/80-net-name-slot.rules which will mask the file of the same name located under /usr/lib/udev/rules.d (alternatively, instead of an empty file, using a symlink to /dev/null is also an acceptable masking method).

Wired

Dynamic IP

Using dhcpcd

If you only use a single fixed wired network connection, you do not need a network management service and can simply enable the dhcpcd service:

# systemctl enable dhcpcd.service

Note: If it doesn't work, use: # systemctl enable dhcpcd@interface_name.service

Using netctl

Copy a sample profile from /etc/netctl/examples to /etc/netctl:

# cd /etc/netctl
# cp examples/ethernet-dhcp my-network

Edit the profile as needed (modify Interface):

# nano my-network

Enable the my-network profile:

# netctl enable my-network

Using netctl-ifplugd

Warning: You cannot use this method in conjunction with explicitly enabling profiles, such as netctl enable <profile>.

Alternatively, you can use netctl-ifplugd, which gracefully handles dynamic connections to new networks:

Install ifplugd, which is required for netctl-ifplugd:

# pacman -S ifplugd

Then enable for interface that you want:

# systemctl enable netctl-ifplugd@<interface>.service

Tip: Netctl also provides netctl-auto, which can be used to handle wired profiles in conjunction with netctl-ifplugd.

Static IP

Using netctl

Copy a sample profile from /etc/netctl/examples to /etc/netctl:

# cd /etc/netctl
# cp examples/ethernet-static my-network

Edit the profile as needed (modify Interface, Address, Gateway and DNS):

# nano my-network

• Notice the /24 in Address which is the CIDR notation of a 255.255.255.0 netmask

Enable above created profile to start it at every boot:

# netctl enable my-network

Wireless

Note: If your wireless adapter requires a firmware (as described in the above Establish an internet connection section and also here), install the package containing your firmware. Most of the time, the linux-firmware package will contain the needed firmware. Though for some devices, the required firmware might be in its own package. For example:

# pacman -S zd1211-firmware

See Wireless Setup#Installing driver/firmware for more info.

Install iw and wpa_supplicant which you will need to connect to a network:

# pacman -S iw wpa_supplicant

Adding wireless networks

Using wifi-menu

Install dialog, which is required for wifi-menu:

# pacman -S dialog

After finishing the rest of this installation and rebooting, you can connect to the network with wifi-menu interface_name (where interface_name is the interface of your wireless chipset).

# wifi-menu interface_name

Warning: This must be done *after* your reboot when you are no longer chrooted. The process spawned by this command will conflict with the one you have running outside of the chroot. Alternatively, you could just configure a network profile manually using the following templates so that you do not have to worry about using wifi-menu at all.

Using manual netctl profiles

Copy a network profile from /etc/netctl/examples to /etc/netctl:

# cd /etc/netctl
# cp examples/wireless-wpa my-network

Edit the profile as needed (modify Interface, ESSID and Key):

# nano my-network

Enable above created profile to start it at every boot:

# netctl enable my-network

Connect automatically to known networks

Warning: You cannot use this method in conjunction with explicitly enabling profiles, such as netctl enable <profile>.

Install wpa_actiond, which is required for netctl-auto:

# pacman -S wpa_actiond

Enable the netctl-auto service, which will connect to known networks and gracefully handle roaming and disconnects:

# systemctl enable netctl-auto@interface_name.service

Tip: Netctl also provides netctl-ifplugd, which can be used to handle wired profiles in conjunction with netctl-auto.

Analog modem, ISDN or PPPoE DSL

For xDSL, dial-up and ISDN connections, see Direct Modem Connection.

Create an initial ramdisk environment

Tip: Most users can skip this step and use the defaults provided in mkinitcpio.conf. The initramfs image (from the /boot folder) has already been generated based on this file when the linux package (the Linux kernel) was installed earlier with pacstrap.

Here you need to set the right hooks if the root is on a USB drive, if you use RAID, LVM, or if /usr is on a separate partition.

Edit /etc/mkinitcpio.conf as needed and re-generate the initramfs image with:

# mkinitcpio -p linux

Note: Arch VPS installations on QEMU (e.g. when using virt-manager) may need virtio modules in mkinitcpio.conf to be able to boot.

# nano /etc/mkinitcpio.conf

MODULES="virtio virtio_blk virtio_pci virtio_net"

Set the root password

Set the root password with:

# passwd

Install and configure a bootloader

For BIOS motherboards

For BIOS systems, two bootloaders are available: Syslinux and GRUB. Choose the bootloader as per your convenience.

• Syslinux is (currently) limited to loading only files from the partition where it was installed. Its configuration file is considered to be easier to understand. An example configuration can be found here.

• GRUB is more feature-rich and supports more complex scenarios. Its configuration file(s) is more similar to a scripting language, which may be difficult for beginners to manually write. It is recommended that they automatically generate one.

Syslinux

Note: If you opted for a GUID partition table (GPT) for your hard drive earlier, you need to install the gptfdisk package now for this next step to work, assuming you have not installed it already.

Install the syslinux package and then use the syslinux-install_update script to automatically install the bootloader (-i), mark the partition active by setting the boot flag (-a), and install the MBR boot code (-m):

# pacman -S syslinux
# syslinux-install_update -i -a -m

Configure syslinux.cfg to point to the right root partition. This step is vital. If it points to the wrong partition, Arch Linux will not boot. Change /dev/sda3 to reflect your root partition (if you partitioned your drive as in the example, your root partition is sda1). Do the same for the fallback entry.

# nano /boot/syslinux/syslinux.cfg

...
LABEL arch
        ...
        APPEND root=/dev/sda3 rw
        ...

For more information on configuring and using Syslinux, see Syslinux.

GRUB

Install the grub package and then run grub-install to install the bootloader:

Note:

• Change /dev/sda to reflect the drive you installed Arch on. Do not append a partition number (do not use sdaX).
• For GPT-partitioned drives on BIOS motherboards, you also need a "BIOS Boot Partition". See GPT-specific instructions and Install to GPT BIOS boot partition in the GRUB page.

# pacman -S grub
# grub-install --recheck /dev/sda

Note: If it is an installation on virtualbox as guest, while running grub-install command as in above, you might get an error like "/usr/sbin/grub-bios-setup: warning: this GPT partition label contains no BIOS Boot Partition; embedding won't be possible". Executing parted -s /dev/sda set 1 bios_grub on and then retrying grub-install should solve the problem.

While using a manually created grub.cfg is absolutely fine, it is recommended that beginners automatically generate one:

Tip: To automatically search for other operating systems on your computer, install os-prober (pacman -S os-prober) before running the next command.

# grub-mkconfig -o /boot/grub/grub.cfg

For more information on configuring and using GRUB, see GRUB.

For UEFI motherboards

For UEFI systems, several options are available. A complete list of options is available at UEFI Bootloaders. You may find that some options work while others do not. Otherwise, choose one as per your convenience. Here, we give two of the possibilities as examples:

• Boot the Linux kernel directly using EFISTUB.
• gummiboot is a simple boot manager, useful if you are dual booting. rEFInd is another alternative.
• GRUB is a more complete bootloader, useful if you run into problems with the other two options.

Note: For UEFI boot, the drive needs to be GPT-partitioned and an UEFI System Partition (512 MiB or larger, type EF00, formatted with FAT32) must be present. For the following examples, this partition must be mounted on /boot. If you have followed this guide from the beginning, you have already done all of these.

Note: If you run into problems, such as not being able to boot after the bootloader has been installed without any visible error. In this case, you will instead have to enter the UEFI shell and manually add an entry to the UEFI boot menu with the bcfg command, as described here.

• On some ASUS motherboards, there is an EFI bug that always reports MaxVariableSize=0. Combined with a recent kernel that enforces checks on this value, this prevents efibootmgr from setting new EFI variables. These motherboards do not support the UEFI Shell v2, so you cannot use the bcfg method either. To work around this, add efi_no_storage_paranoia to the kernel command line. You can do this by pressing "e" at the bootloader menu.
• On some UEFI motherboards like the Intel Z77 boards, adding entries with efibootmgr or bcfg from efi shell will not work because they don't show up on the boot menu list after being added to NVRAM.

To solve this you have to trick the UEFI firmware that Windows boot manager is present on the ESP partition.

Copy the bootx64.efi file from USB drive as bootmgfw.efi efi file to your ESP partition by booting into EFI shell and typing:

FS1:
cd EFI
mkdir Microsoft
cd Microsoft
mkdir Boot
cp FS0:\EFI\BOOT\bootx64.efi FS1:\EFI\Microsoft\Boot\bootmgfw.efi

After reboot, any entries added to NVRAM should show up in the boot menu.

EFISTUB

Install the efibootmgr package and then add an Arch Linux boot entry, replacing /dev/sdaX with your root partition, usually /dev/sda2:

# pacman -S efibootmgr
# efibootmgr -c -L "Arch Linux" -l /vmlinuz-linux -u "root=/dev/sdaX rw initrd=/initramfs-linux.img"

Gummiboot

Install the gummiboot package and then run gummiboot install to install the boot manager:

# pacman -S gummiboot
# gummiboot install

Warning: You will probably see an error during gummiboot install, when it fails to add itself to NVRAM because of a bug. If you get this error message, manually use efibootmgr to add gummiboot to NVRAM:

# efibootmgr -c -L "Gummiboot" -l /EFI/gummiboot/gummibootx64.efi

You will need to manually create a configuration file to add an entry for Arch Linux to the gummiboot manager. Create /boot/loader/entries/arch.conf and add the following contents, replacing /dev/sdaX with your root partition, usually /dev/sda2:

# nano /boot/loader/entries/arch.conf

title          Arch Linux
linux          /vmlinuz-linux
initrd         /initramfs-linux.img
options        root=/dev/sdaX rw

For more information on configuring and using gummiboot, see gummiboot.

GRUB

Install the grub and efibootmgr packages and then run grub-install to install the bootloader:

# pacman -S grub efibootmgr
# grub-install --efi-directory=/boot --bootloader-id=arch_grub --recheck

Next, while using a manually created grub.cfg is absolutely fine, it is recommended that beginners automatically generate one:

Tip: To automatically search for other operating systems on your computer, install os-prober (pacman -S os-prober) before running the next command.

# grub-mkconfig -o /boot/grub/grub.cfg

For more information on configuring and using GRUB, see GRUB.

Unmount the partitions and reboot

Exit from the chroot environment:

# exit

Since the partitions are mounted under /mnt, we use the following command to unmount them:

# umount /mnt/{boot,home,}

Reboot the computer:

# reboot

Tip: Be sure to remove the installation media, otherwise you will boot back into it.