Seconding this. Store your configuration.nix in git and just copy it back over if you ever need to wipe and reinstall.
Git.
Keep all the config files of your tools in subdirectories of a git versioned directory and symlink them into their target location (e.g. with GNU stow). If installation of a tool is involved and you expect to have to revisit it, put the steps into an installation bash script and version it as well.
+1, essential ones I keep in GitHub repository (like zsh, tmux, xdefaults configs with no personal data). With makefile that makes symlinks. This is the easiest way to sync zsh config between my personal and work machines.
Rest is just in a backup.
Do you have an example of a generalise makefile that does that? Or does it need to be customise per configuration?
On my GitHub repo. Needs to be customized, but you should get the idea.
Maybe there is a way to write it better, I'm no makefile expert ;)
@zacher_glachl @perishthethought I take a similar approach starting with a bare work-tree at $HOME/.cfg
and add config files I've changed. Then throw my --git-dir
and --work-tree
switches in an alias for git.
As for installed programs, a simple backup of my portage world file takes car of that.
Oh! I can participate!
Everything I have/configure is 100% in Ansible. I learned the hard way that rebuilding a workstation from scratch sucks if I only depend on my brain to remember things.
It takes some effort to keep it updated - if I'm trying out a new app, I have to remember to add it to my config.
The other thing that I've started doing is using Restic for file level backups. That's relatively easy to set up, it supports a multitude of backend storage, and works well with a cron job for braindead backups.
Home Manager on a NixOS flake, it's a rabbit hole but I've been loving it since last week!
I just check my Nix.config, but most distros don't have that privilege.
Idk how it works for most other distros, but I know on Arch you can check all packages manually installed by pacman and your AUR helper.
Iirc pacman -Qe
does something like this
Install fresh copy of Linux OS on a new device. Install the apps I know I need like browser, code editor, etc.
Use device.
Realize "oh crap I forgot to install X!"
Install X
Repeat until all X have been installed.
Lol, pretty much what I've done in the past. And yep, it works, eventually.
Plus backup /home
and /etc
, maybe /opt
as well.
Move all your heavily modified config files into a git repository and host it somewhere. Then symlink all your config files to where they should be with ln -s ~/.config/whatever ~/gitrepo/whatever
. That's how you preserve your important configs.
You can easily get a list of your installed packages (which you can keep in your repository) with apt list --installed > packages.txt
. You can then format that list to one you can install from with sed -e "s-/.*$--" <packages.txt
(or something, i don't have apt, can't test it fully).
In fact, if someone here is more familiar with apt, please find a way to filter out packages that were not explicitly installed and reply to this comment with your solution.
This combined with stow
command makes it very simple to "install" your system configuration on a new machine.
Apart from those little tools running in the background that get their own little "How to install and configure X" file, I don't keep a list. I just install things as I need them, copying back config files from a backup. It's less annoying and time consuming than one might expect and keeps the system slim by not installing what I never use anyway.
Backup $home and /etc. That should be good enough.
Yes, my concern with this is that I have Steam installed and its games are many many GBs. If I do a backup I'll have to exclude that folder. I'll try this and see how it goes. Thanks!
Use ansible and variable, so it can be replicated to other computer. Simple
ansible.builtin.apt:
name: "{{ item }}"
state: latest
loop:
- pkg1
- pkg2
At work we are starting to transition to Ansible from chef and other homegrown solutions. So to learn Ansible I added awx to my home lab and now have playbooks for almost all of my devices. Going to format a Pi again soon and see if everything works as intended
git repo
A bash script
apt get install
Same with flatpak
Keep updated. Done.
I make a list of all the ones I like. Then when I feel my system is getting too bloated, I wipe and reinstall while only installing the packages from my list.
It’s very “low tech,” but it’s always worked out well for me.
Yes, I think this is more like what I'll do, though I like the idea of a git repo for the configurations. Cheers
How do you keep track of what you’ve installed/ your favorite apps?
https://github.com/Atemu/nixos-config/blob/b79f42793a709db083cf53867f85d5d46e41eb69/packages.nix
Separately, how can I backup the configurations I’m using right now.
NixOS stores a snapshot of your OS and all the app configs in an OS config folder for you. Helpful for instant system recovery or deploying the setup to new hardware.
https://github.com/koepnick/dotfiles cloned into ~/.config
I typically start with a restrictive .gitignore and add directories as needed.
A ton of stuff that I always forget like mpv, vifm, and whatnot always slipped through the cracks before. Now I can clone to practically anywhere and have everything just work.
Lately using Silverblue. Everything is a flatpak or is layered. Both are easy to list.
Maybe a handful of things in distrobox I need to keep track of.
If you use other package managers like flatpak
, nix-env
(for non-NixOS), npm
(for global stuff), … you can just create a bash script to list all installed packages for each manager, and save to a file(s). I put them in a git repo to version control these lists, which are updated every now and then when I update stuff.
I keep a changelog for all my computers and note installations, uninstallations and config changes.
I use a notepad and write down the apps I mostly use so when I do a reinstall I just reference that.
A dot files repo for some basic config and an Ansible repo to stand everything up. This applies to both my Linux and MacOS machines.
All the people taking about their NixOS setups had me thinking of giving that a try, though.
I back up the entire system, apps and all. The main reason is so that I can get back up and running quickly, but it also spares me the need to worry about which apps to reinstall.
😂😂😂 😭😭😭😭
Ansible, although that may be overkill.
I personally considered doing this but the time i would have to invest vs the frequency of doing a fresh install just doesn't warrant it. Too complex.
I screw around with self hosting and homelab stuff that it was an interesting thing to learn at a certain point.
I don't think I noticed which c this was in. For a regular user it's probably not worth the effort.
Well i write frequently write playbooks at work. But for my laptop it Still felt not worth it. I have a playbook for my Server stuff though.
function pkglist -d "Gets list of installed packages"
# Prevent descriptions in other languages
set -l LANG C
# Define pkglists location
set -l dot $HOME/.config/dotfiles
echo "(1/5) RPM-OSTREE status"
rpm-ostree status > $dot/pkglist.rpm-ostree --booted
echo "(2/5) Identify flatpaks"
flatpak list --app --columns=application > $dot/pkglist.flatpak
echo "(3/5) Identify pinned flatpak runtimes"
flatpak pin > $dot/pkglist.flatpak.pinned
echo "(4/5) Identify flatpak overrides"
for i in (cat $dot/pkglist.flatpak)
if test -s (flatpak override --show --user $i|psub)
echo $i
flatpak override --show --user $i
echo
end
end > $dot/pkglist.flatpak.overrides
echo "(5/5) Save KDE configuration"
fedora konsave -s kde_configuration --force
# to apply configuration
# fedora konsave -a kde_configuration
git -C $dot st
end
Write install and maintenance guides, save them somewhere public. Automation is good, but documentation and practicing technical writing is better.
I keep a .dotfiles folder in my home dir, use syncthing to back up those files on a couple of other computers, and then (on a new install) just make the actual config files symlinks to those files.
Linux
From Wikipedia, the free encyclopedia
Linux is a family of open source Unix-like operating systems based on the Linux kernel, an operating system kernel first released on September 17, 1991 by Linus Torvalds. Linux is typically packaged in a Linux distribution (or distro for short).
Distributions include the Linux kernel and supporting system software and libraries, many of which are provided by the GNU Project. Many Linux distributions use the word "Linux" in their name, but the Free Software Foundation uses the name GNU/Linux to emphasize the importance of GNU software, causing some controversy.
Rules
- Posts must be relevant to operating systems running the Linux kernel. GNU/Linux or otherwise.
- No misinformation
- No NSFW content
- No hate speech, bigotry, etc
Related Communities
Community icon by Alpár-Etele Méder, licensed under CC BY 3.0