Leaving arch for debian or red hat distros

I am facing a lot of challenges in terms of installing crucial data science packages, such as OpenFoam, Tensorflow with GPU support, docker, etc. Of course, it still works with the help of AUR and what not, but it seems like a hack and eventually I don’t really get the official set-up.

Am I missing something? What would you suggest me to do in this case? I really like EndevourOS and I would really like to stay with it, but I keep running into these issues. :frowning: It seems like Arch is very underrated in the Data Science community.

1 Like

I’m not entirely sure I understand your question. Are you saying you have actual software issues in EndeavourOS or are you saying you have ideological issues with using the AUR?

For example right now I’m trying to install OpenFoam and it’s compiling it for the last 40 minutes, whereas OpenFoam and other similar packages in Data Science are built mainly for Ubuntu and Fedora and the installation takes seconds. I am trying to understand if this is what I get with arch or am I missing something by doing things in the wrong way. I am very new to Linux.

may have a look into communities specialized on such things:
https://wiki.archlinux.org/title/Unofficial_user_repositories#arch4edu

3 Likes

I’m not the best expert in explaining the differences with AUR things. For that I’m sure @dalto could chip in, but I’ll briefly offer my own experience. Depending on the program in the AUR, some will build quickly and some will take 30 minutes or even longer. Part of that depends on your internet speed, your computers hardware, and what all is needed to be built. For example, installing vivaldi-snapshot in the AUR takes a few quick minutes to build. But for video/audio codec support, you have to also build vivaldi-snapshot-ffmpeg-codecs, and that takes easily 30-40 minutes to build.

Package types you’ll see in the AUR:

  • Normal packages are built from stable versions or stable git tags of a repository. The program is compiled in user’s machine and then installed. This will take time.
  • Packages with -bin suffix are already built by upstream maintainer and is available somewhere. So, users do not have to compile the package in their machine. The PKGBUILD script downloads, extracts and install the files. Some proprietary software are released in this format where source code is not available.
  • Packages with -git suffix are built from the latest commit from git repository, no matter it is a stable or not. This way user get latest fix or patches. This also compiled in user machine, then installed.

Source: click here

In the AUR, you have a few different package versions to choose based on what you prefer. You’ll have the main package, git package, a bin package as mentioned above. A bin package is already built so the process is a lot easier/quicker, so that would be what I’d suggest for you. However, for OpenFoam there is not bin package available in the AUR, so you’ll have to always build it yourself. IF you can find any other packages that you use in the AUR for this type, I would suggest using the bin packages and see if that helps.

But in all honesty, if building packages and the time it may sometimes require to make is not your cup of tea, there are alternatives for data science distros, mainly Ubuntu or Fedora. I would try bin packages if and when you can. If you still don’t fancy the long build updates, you may have to go on a little distro hopping journey and there’s nothing wrong with that if you’re able to find a distro what works better for your use case. I hope this was helpful, as it helped remind me what the AUR is all about :wink:

2 Likes

This was super helpful. Thank you very much for the detailed response.

1 Like

From the link @joekamprad provided from the Arch wiki, if you visit the Arch4edut github:

it lists all the programs they currently have in their repo. Hopefully you can find all the ones you need within there. IF not, you can always open up an issue to request a program you need be included:

I don’t have any experiencing adding, maintaining, or updating unofficial user repositories, but I’m sure some users here have the experience and knowledge to help you out with that if you get stuck.

2 Likes

I’m considering this option. So I have added the mirror to my /etc/pacman.conf but I’m not sure what I should do next. Would you please guide me on that?

Can you show your /etc/pacman.conf for a reference to see what exactly you added?

here is what it looks like if I run sudo nano /etc/pacman.conf.

GNU nano 5.9     /etc/pacman.conf.                                                 
[arch4edu]

Server = https://arch4edu.keybase.pub/$arch

this is not looking good to me… if it is all inside the config…

cat /etc/pacman.conf | eos-sendlog will pastebin it and give URL to the pastebin :wink:

Please tell me that’s not your entire pacman.conf??

1 Like

It is my entire pacman.conf. Is my PC going to explode soon? :innocent:

https://clbin.com/7HGa8
should look like so

################################################################################
################# Arch Linux mirrorlist generated by Reflector #################
################################################################################
# With:       reflector --verbose -c US --protocol https --sort rate --latest 10
# When:       2021-11-08 02:52:11 UTC
# From:       https://archlinux.org/mirrors/status/json/
# Retrieved:  2021-11-08 02:51:23 UTC
# Last Check: 2021-11-08 02:25:28 UTC

## United States
Server = https://mirrors.ocf.berkeley.edu/archlinux/$repo/os/$arch

## United States
Server = https://arch.hu.fo/archlinux/$repo/os/$arch

## United States
Server = https://mirrors.rit.edu/archlinux/$repo/os/$arch

## United States
Server = https://mirror.dal10.us.leaseweb.net/archlinux/$repo/os/$arch

## United States
Server = https://mirror.wdc1.us.leaseweb.net/archlinux/$repo/os/$arch

## United States
Server = https://mirror.mia11.us.leaseweb.net/archlinux/$repo/os/$arch

## United States
Server = https://mirror.lty.me/archlinux/$repo/os/$arch

## United States
Server = https://mirror.sfo12.us.leaseweb.net/archlinux/$repo/os/$arch

## United States
Server = https://mirrors.lug.mtu.edu/archlinux/$repo/os/$arch

## United States
Server = https://repo.ialab.dsu.edu/archlinux/$repo/os/$arch

they have a tutorial how to add the repo also

After this you can search and install the packages from the repo with pacman.

That’s actually your pacman mirrorlist, which is different from /etc/pacman.conf, FYI

I did that. Probably that wiped my config?

  1. Then add
[arch4edu]
Server = <the mirror site you choose>

to your /etc/pacman.conf.

they do not say anything like replace it says add :wink:

Right when I opened the config it was already empty I swear lol. How do I get back the config now?