Offline installation of packages
From time to time, you might end up with a copy of Arch Linux on an offline computer and want to install packages for it.
Install from file
See pacman#Additional commands to learn how to install local packages. Shell globbing can be used to install many packages at once:
# pacman -U /package/folder/*.tar.zst
Offline cache
You can put things into /var/lib/pacman/sync
and /var/cache/pacman/pkg
so as to make pacman
think it has everything it needs to do searches, updates, and installs. This method is based on byte's post from this thread.
Download the package databases on a computer with internet access and transfer them to your computer. If needed, change MIRROR
to any mirror from the mirror status list.
You might have changed your default repositories from its default order (community, core, extra and multilib); if you did this in your specific installation, you should review your /etc/pacman.conf
file.
#!/bin/sh ARCH="x86_64" MIRROR="https://mirrors.kernel.org/archlinux/" wget "${MIRROR}/community/os/${ARCH}/community.db" wget "${MIRROR}/core/os/${ARCH}/core.db" wget "${MIRROR}/extra/os/${ARCH}/extra.db" wget "${MIRROR}/multilib/os/${ARCH}/multilib.db" # and possibly -uncomment- (if customized in /etc/pacman.conf or pacman.conf.d): #wget "${MIRROR}/multilib-testing/os/${ARCH}/multilib-testing.db" #wget "${MIRROR}/community-testing/os/${ARCH}/community-testing.db" #wget "${MIRROR}/testing/os/${ARCH}/testing.db" # and -additionaly- debug and staging packages.
The following steps will make sure you are working with up-to-date package lists, as if you ran .
After transferring the files to the offline PC, do the following:
# cp *.db /var/lib/pacman/sync/ # pacman -Sp --noconfirm package-name > pkglist
To update a New Arch Linux base system after installation, you may enter:
# pacman -Sup --noconfirm > pkglist
Now, open that text file with an editor and delete all lines that are not URLs.
Next, bring that list with you to a place where you have internet and either download the listed packages manually or run wget
in an empty directory:
# wget -nv -i ../pkglist
You must also download the corresponding package signatures:
# sed -e 's/\.zst$/.zst.sig/' ../pkglist > ../siglist # wget -nv -i ../siglist
wget --restrict-file-names=unix
.Take all the and files back home, put them in /var/cache/pacman/pkg
and finally run
# pacman -S package-name
Local repository
Scenario: you have two Arch Linux machines, 'Al' (with internet connection) and 'Bob' (without internet connection), and you need to install some NVIDIA packages and their dependencies on 'Bob'. In this example, the wanted packages are , , and , but you want to use a dedicated directory instead of and a dedicated repository called [nvidia] (instead of the usual [core], [extra] etc...)
Generate a list of packages to download
This can be done on any Arch Linux machine which has up-to-date repository data bases (see above for links to database files); to create the list of links to the required packages, use:
# pacman -Sp nvidia nvidia-utils xf86-video-nouveau > /path/to/nvidia.list
The file will contain links to the listed packages and any others which they depend on which are not already installed on 'Al'. Unless you have cleared your cache, the packages you have installed will be in your cache location. You can check /etc/pacman.conf
for the location. It is probably something like .
Download/copy the packages and their dependencies
Obviously, this requires an internet connection, so on 'Al', create a directory called for the files and run:
# wget -P /path/to/nvidia/ -i /path/to/nvidia.list
Then copy the dependencies you have already installed from the cache. Either find them manually by browsing https://archlinux.org/packages/ or if the total size of all your packages is not too large, just copy them all
# cp /var/cache/pacman/pkg/* /path/to/nvidia/
Create a repository database just for these packages
This can be done on either 'Al' or 'Bob' using the repo-add
command which comes with (from version 3?); first, change to the directory where the packages were downloaded, then create database file called nvidia.db.tar.gz
:
$ cd /path/to/nvidia # repo-add nvidia.db.tar.gz *.pkg.tar.zst
Transfer the packages
Now that all the packages have been downloaded, you do not need 'Al' anymore. Copy the contents of to a the temporary NVIDIA packages cache directory on 'Bob'. In this example, this folder is called :
$ cp /path/to/nvidia/* /home/me/nvidia
Next, must be made aware of this new repository of packages. First, copy your current :
# cp /etc/pacman.conf /etc/pacman.conf.old
Now, in /etc/pacman.conf
, make sure that your is set to as your repository will not provide signatures
SigLevel = Never
and add the following lines at the bottom of :
[nvidia] Server = file:///home/me/nvidia
You may also need to comment out the other repositories so stale defaults do not cause failed attempts to download from online Now, instruct to synchronize with the dedicated NVIDIA repository we created:
# pacman -Sy
This command finds the nvidia.db.tar.gz
file in and expands it to to create a database of packages contained in the NVIDIA repository.
Install the packages
Finally, install the packages:
# pacman -S nvidia nvidia-utils xf86-video nouveau
Restoring online sources
Should Bob ever be put online, we can restore access to the online sources by replacing /etc/pacman.conf
with the previously created .
Links and sources
Compiled from the forums, with thanks to Heller_Barbe and byte