Beginners' guide

From ArchWiki
Revision as of 00:08, 8 April 2013 by KevinCoghlan (Talk | contribs) (Clarified wording.)

Jump to: navigation, search
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.


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, 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 Template:Keypress to exit, and when prompted to save changes, press Template:Keypress and Template:Keypress 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, Template:Keypress activates and deactivates the keymap.

Establish an internet connection

Warning: udev no longer assigns network interface names according to the wlanX and ethX naming scheme. If you're 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 addr show to discover the names of your interfaces.

From systemd-197's release and onward, udev now assigns predictable, stable network interface names that deviate from the legacy incremental naming scheme (wlan0, wlan1, etc.). These interface names are guaranteed to be persistent across reboots, which solves the problem of the lack of predictability of network interface name assignment. For more information about why this was necessary, read .

The dhcpcd network daemon is started automatically at boot and it will attempt to start a wired connection, if available. Try pinging a website to see if it was successful. And since Google is always on...

# ping -c 3
PING ( 56(84) bytes of data.
64 bytes from ( icmp_req=1 ttl=50 time=17.0 ms
64 bytes from ( icmp_req=2 ttl=50 time=18.2 ms
64 bytes from ( icmp_req=3 ttl=50 time=16.6 ms

--- 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 any problem with your cable (or if you have enough wireless signal), otherwise you will need to set up the network manually, as explained below.

Otherwise, move on to Prepare the storage drive.


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

First, 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're 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're 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 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

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

# nano /etc/resolv.conf
nameserver 61.95.849.8
Note: Currently, you may include a maximum of three nameserver lines.

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


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

First, identify the name of your wireless interface.

# 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, wlp3s0 is the available wireless interface. If you're 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 the wireless chipset requires firmware, you are likely to receive this error when bringing the interface up:

# ip link set wlp3s0 up
SIOCSIFFLAGS: No such file or directory

If unsure, invoke dmesg to query the kernel log for a firmware request from the wireless chipset.

Example output from 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 netcfg'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.

xDSL (PPPoE), analog modem or ISDN

If you have a router in bridge mode, run:

# pppoe-setup

To use these settings and connect to your ISP, run:

# pppoe-start

You may also need to adjust your resolv.conf:

# echo nameserver > /etc/resolv.conf

If you have a dial-up or ISDN connection, 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.

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. Also PartedMagic is a live distribution that can be booted either from CD/DVD or a usbkey and has Gparted as the graphical partition facility.

The recommendation for a system that will boot via UEFI rather than MBR legacy boot is to format the drive using a GPT partition table. This means that if the drive was previously already partitioned with an MBR (MSDOS) partition table it will now have a new partition table created which will destroy all other data on the drive. Once the new partition table has been created on a drive, only then can individual partitions be created with any chosen format type. 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 the advanced option, and then select a GPT partition table. This can't be done if you have a pre-existing Windows installation on the drive which you wish not to destroy. It is therefore extremely important that if you intend to have a dual boot system and not change the partition table to GPT, you need to be aware of the pitfalls of leaving the Windows install untouched and trying to get the linux install working with UEFI on the drive that remains with an MBR (legacy) partition table.

In addition 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 then 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 even for linux distributions that support it is still not in a fully stable state for reliable operation.

See Swap for details if you wish to set up a swap partition or file now. 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.

If you have already done so, proceed to Mount the partitions.

Otherwise, see the following example.


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

Notes regarding UEFI boot:
  • If you have a UEFI motherboard, you will need to create an extra UEFI System Partition.
  • It is recommended to always use GPT for UEFI boot, as some UEFI firmwares do not allow UEFI-MBR boot.
Notes regarding GPT partitioning:
  • If you are not dual booting with Windows, then it is advisable to use GPT instead of MBR. Read GPT for a list of advantages.
  • 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. Syslinux doesn't need one.
  • Some BIOS systems may have issues with GPT. See and for more info and possible workarounds.
Note: If you are installing to a USB flash key, see Installing Arch Linux on a USB key.

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.

MBR cfdisk /dev/sda Root:


GPT cgdisk /dev/sda Root:

If you chose MBR, here's how it should look like:

Name    Flags     Part Type    FS Type          [Label]       Size (MB)
sda1    Boot       Primary     Linux                             15360
sda2               Primary     Linux                             133000*

If you chose GPT, here's how 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 Template:Keypress) to exit without saving changes and then restart cfdisk (or cgdisk).

If you are satisfied, choose Write (or press Template:Keypress) to finalize and to write the partition table to the drive. Type "yes" and choose Quit (or press Template:Keypress) to exit without making any more changes.

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's actually /dev/sda1 and /dev/sda2 that you want to format.
# mkfs.ext4 /dev/sda1
# mkfs.ext4 /dev/sda2

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

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

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 when using cfdisk 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/efi
# mount /dev/sdaX /mnt/boot/efi

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's 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 =$repo/os/$arch

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

  • 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.
  • 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're 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 and base-devel groups without prompting.

# pacstrap -i /mnt base base-devel
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.
Note: If pacman complains that error: failed to commit transaction (invalid or corrupted package), run the following command:
# pacman-key --init && pacman-key --populate archlinux
  • base: Software packages from the [core] repo to provide the minimal base environment.
  • base-devel: Extra tools from [core] such as make, and automake. Most beginners should choose to install it, as it will likely be needed to expand the system. The base-devel group will be required to install software from the Arch User Repository.

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.

Note: If you encounter errors running genfstab or later in the install process, do not run genfstab again; just edit the fstab file.
# genfstab -U -p /mnt >> /mnt/etc/fstab
# nano /mnt/etc/fstab

A few considerations:

  • Only the root (/) partition needs 1 for the last field. Everything else should have either 2 or 0 (see fstab#Field definitions).

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.


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 doesn't 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 – 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's 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


# 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 don't 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.


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.


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.


Dynamic IP
Warning: A bug has been noted in the install ISO, in which the name your interface has during installation differs from the one it will have upon reboot. See Bug #33923 for more details.
Use the command ip link (shows interface names) after rebooting into your installed system to find out if you are affected by this. If so, you will have to redo the configuration described below with the correct interface name.

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. Here, interface_name is your wired interface:

# systemctl enable dhcpcd@interface_name.service

And that's it.

Move on to the next step: Create an initial ramdisk environment.

Alternatively, you can use netcfg's net-auto-wired, which gracefully handles dynamic connections to new networks:

Install ifplugd, which is required for net-auto-wired:

# pacman -S ifplugd

Edit /etc/conf.d/netcfg and modify the network interface name, most likely it is not eth0. You can find out more about the naming in the warning above.

# nano /etc/conf.d/netcfg

Copy a sample profile from /etc/network.d/examples to /etc/network.d:

# cd /etc/network.d
# cp examples/ethernet-dhcp .

Edit the profile as needed (modify INTERFACE):

# nano ethernet-dhcp

Enable the net-auto-wired service:

# systemctl enable net-auto-wired.service
Static IP

Copy a sample profile from /etc/network.d/examples to /etc/network.d:

# cd /etc/network.d
# cp examples/ethernet-static .

Edit the profile as needed (modify INTERFACE, ADDR, GATEWAY and DNS):

# nano ethernet-static

Edit /etc/conf.d/netcfg and add the new network profile to the NETWORKS array:

nano /etc/conf.d/netcfg

Enable the netcfg service:

# systemctl enable netcfg.service


You will need to install additional programs to be able to configure and manage wireless network profiles for netcfg.

NetworkManager and Wicd are other popular alternatives.

  • Install the required packages:
# pacman -S wireless_tools wpa_supplicant wpa_actiond dialog

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. For example:

# pacman -S zd1211-firmware

See Wireless Setup and WPA supplicant for more info.

  • 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), which will generate a profile file in /etc/network.d named after the SSID. There are also templates available in /etc/network.d/examples/ for manual configuration.
# wifi-menu interface_name
Warning: If you're using wifi-menu, this must be done *after* your reboot when you're 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 templates previously mentioned so that you don't have to worry about using wifi-menu at all.
  • Enable the net-auto-wireless service, which will connect to known networks and gracefully handle roaming and disconnects:
# systemctl enable net-auto-wireless.service
Note: From Netcfg#Net-Auto-Wireless: wireless-wpa-config profiles do not work with net-auto-wireless. Convert them to wireless-wpa-configsection or wireless-wpa instead.
Note: Netcfg also provides net-auto-wired, which can be used in conjunction with net-auto-wireless.
Note: Wpasupplicant could be fail with message "WPA Authentication/Association Failed". In that case, see this link for a solution.
  • Make sure that the correct wireless interface (e.g. wlp3s0) is set in /etc/conf.d/netcfg:
# nano /etc/conf.d/netcfg

It is also possible to define a list of network profiles that should be automatically connected, using the AUTO_PROFILES variable in /etc/conf.d/netcfg. If AUTO_PROFILES is not set, all known wireless networks will be tried.

xDSL (PPPoE), analog modem or ISDN

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, there are three bootloaders - Syslinux, GRUB, and LILO. Choose the bootloader as per your convenience. Below only Syslinux and GRUB are explained.

  • 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.
Note: Some BIOS systems may have issues with GPT. See and for more info and possible workarounds.

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

Note: If you have partitioned the drive as GPT, install gptfdisk package, as well (pacman -S gptfdisk), because it contains sgdisk, which will be used to set the GPT-specific boot flag.
# 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 ro

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


Install the grub-bios package and then run grub-install /dev/sda:

Note: Change /dev/sda to reflect the drive you installed Arch on. Do not append a partition number (do not use sdaX).
Note: For GPT-partitioned drives on BIOS motherboards, GRUB needs a "BIOS Boot Partition".
# pacman -S grub-bios
# grub-install --recheck /dev/sda
# cp /usr/share/locale/en\@quot/LC_MESSAGES/ /boot/grub/locale/

While using a manually created grub.cfg is absolutely fine, it's 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 GRUB2.

For UEFI motherboards

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

While there are other UEFI bootloaders available, using EFISTUB is recommended. Below are instructions for setting up EFISTUB and GRUB (of course you choose only one of them).

Note: Syslinux does not yet support UEFI.

The Linux kernel can act as its own bootloader using EFISTUB. This is the UEFI boot method recommended by developers and simpler compared to grub-efi-x86_64. The steps below set up rEFInd to provide a menu for EFISTUB kernels, as well as for booting other UEFI bootloaders. Alternative EFISTUB boot managers can be found on the page UEFI Bootloaders#Booting EFISTUB. Both rEFInd and gummiboot can detect Windows UEFI bootloaders in case of dual-boot.

1. Boot in UEFI mode and load efivars kernel module before chrooting:

# modprobe efivars      # before chrooting

2. Mount the UEFI System Partition on /mnt/boot/efi, chroot and copy the kernel and initramfs files to /mnt/boot/efi:

  • Create /boot/efi/EFI/arch/ directory.
  • Copy /boot/vmlinuz-linux to /boot/efi/EFI/arch/vmlinuz-arch.efi. The .efi file extension is very important as some UEFI firmwares refuse to launch a file without this extension. Important: Remember that the file is called vmlinuz, not vmlinux.
  • Copy /boot/initramfs-linux.img to /boot/efi/EFI/arch/initramfs-arch.img.
  • Copy /boot/initramfs-linux-fallback.img to /boot/efi/EFI/arch/initramfs-arch-fallback.img.

Every time the kernel and initramfs files are updated in /boot, they need to be updated in /boot/efi/EFI/arch. This can be automated using systemd.

3. For the rEFInd boot manager, install the following packages:

# pacman -S refind-efi efibootmgr

4. Install rEFInd to the UEFI System Partition (summarized from UEFI Bootloaders#Using rEFInd):

# mkdir -p /boot/efi/EFI/refind
# cp /usr/lib/refind/refind_x64.efi /boot/efi/EFI/refind/refind_x64.efi
# cp /usr/lib/refind/config/refind.conf /boot/efi/EFI/refind/refind.conf
# cp -r /usr/share/refind/icons /boot/efi/EFI/refind/icons

5. Create a refind_linux.conf file with the kernel parameters to be used by rEFInd:

# nano /boot/efi/EFI/arch/refind_linux.conf
"Boot to X"          "root=/dev/sdaX ro rootfstype=ext4"
"Boot to console"    "root=/dev/sdaX ro rootfstype=ext4"
Note: refind_linux.conf is copied in the directory /boot/efi/EFI/arch/ where the initramfs and the kernel have been copied to in step 2.
Note: In refind_linux.conf, sdaX refers to your root file system, not your boot partition, if you created them separately.

6. Add rEFInd to UEFI boot menu using efibootmgr.

Warning: Using efibootmgr on Apple Macs may brick the firmware and may need reflash of the motherboard ROM. For Macs, use mactel-bootAUR, or "bless" from within OS X.
# efibootmgr -c -g -d /dev/sdX -p Y -w -L "rEFInd" -l '\EFI\refind\refind_x64.efi'
Note: In the above command, X and Y denote the drive and partition of the UEFI System Partition. For example, in /dev/sdc5, X is "c" and Y is "5".
Note: On some systems, the above command will not work properly. It will execute without any visible error, but the UEFI boot menu will not have been correctly updated with a new entry. To determine whether the command executed properly, enter 'efibootmgr' without any arguments and see if a new entry has appeared in the list displayed. If there is no new entry, then it will not be possible to enter rEFInd upon reboot, as the UEFI boot menu has been left unchanged. 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: Unified_Extensible_Firmware_Interface#bcfg

7. (Optional) As a fallback, in case efibootmgr created boot entry does not work, copy refind_x64.efi to /boot/efi/EFI/boot/bootx64.efi as follows:

# cp -r /boot/efi/EFI/refind/* /boot/efi/EFI/boot/
# mv /boot/efi/EFI/boot/refind_x64.efi /boot/efi/EFI/boot/bootx64.efi
Note: In case you have a system with 32-bit EFI, like pre-2008 Macs, install grub-efi-i386 instead.
# pacman -S grub-efi-x86_64 efibootmgr
# grub-install --efi-directory=/boot/efi --bootloader-id=arch_grub --recheck
# cp /usr/share/locale/en\@quot/LC_MESSAGES/ /boot/grub/locale/

The next command creates a menu entry for GRUB in the UEFI boot menu. However, as of grub-efi-x86_64 version 2.00, grub-install tries to create a menu entry, so running efibootmgr may not be necessary. See UEFI#efibootmgr for more info.

Warning: Using efibootmgr on Apple Macs may brick the firmware and may need reflash of the motherboard ROM. For Macs, use mactel-bootAUR, or "bless" from within OS X.
# efibootmgr -c -g -d /dev/sdX -p Y -w -L "Arch Linux (GRUB)" -l '\EFI\arch_grub\grubx64.efi'
Note: In the above command, X and Y denote the drive and partition of the UEFI System Partition. For example, in /dev/sdc5, X is "c" and Y is "5".

Next, while using a manually created grub.cfg is absolutely fine, it's 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.

Template:Beginners' Guide navigation