Reinstalling Windows hasn’t been fun since Windows 7. The OS already has most drivers and automatically downloads everything else, I miss skimming through pages of drivers to find the correct one.
You’re telling me that making a Microsoft account isn’t fun? It’s truly a process I look forward to. Cortana is literally the friendliest AI waifu assistant I could ever ask for, how can I say no when she asks me to give up my privacy?
Creating a Microsoft account is as easy as “next next done”, where’s the fun in that?.
Cortana was great on Windows Phone (mostly because I love Jen Taylor’s voice), but they kept taking away features and was basically useless on the desktop which, imho, has no use for an assistant.
Been there, started scripting with PowerShell to have an after-setup-script to change reg entries, install tools (mostly through choco), run them (like dism++ cleanup and o&o shutup10 privacy tweaks) and migrate data from backups. Setting up and migrating took me usually 3-4h of work, sometimes more. With the scripts it’s just: Install Windows, update, reboot, update, reboot, run the script, reboot. Done. It’s like 30min of work.
Good thing I changed to linux, cause you can automate the whole process and preseed it into a debian image or kickstart rpm-based distros. It’s possible to do customized Windows images too. I have tried a lot of times. It never worked like it was supposed to.
You don’t have to do this, I manage some machines that haven’t been reinstalled for over a decade. It’s really just because “it feels cleaner”, I guess.
I’ve been running the same installation of Manjaro since 2018, across three different machines. Each time I’ve upgraded hardware I just pop the SSD out and stick it in the new motherboard. Zero instability or troubles from that. Meanwhile I’ve done that to my wife’s Windows PC and it resulted in going through a whole rigmarole with calling Microsoft because the OS install was suddenly no longer activated.
Linux didn’t even care that I went from AMD to Intel to AMD.
I have 3 clones of my 10yo Manjaro desktop install running on other hardware around my network, including a Proxmox VM. It just jumps across, fires up and I fix the hostname, good to go.
Cloning a base image and creating VMs from it is one of the coolest things. I do it for my VMs on my Proxmox cluster any time i need a new server for something - and yeah just copying my dev desktop to my new laptop for going to a conference was such a great way to avoid hours of setup
It is nowhere near necessary to reinstall the OS to fix anything… at least for Mint and Raspbian which are the two I’ve used over the last decade. I may have done an upgrade on mint a few times. Otherwise it chugged on merrily.
PS: now that I think about it I’ve never reinstalled windows on my old laptop either. I like to find the root cause of problems and fix them rather than giving up and reinstalling… call me crazy?
I reinstall about every 6 months, or whenever there is a big feature update. It’s rather noticeable when running benchmarks that performance drops over time mostly 0.1% lows.
Especially when running a stripped install, Microsoft somehow always finds a way to enable shit again or reinstall bloat with updates.
This, plus I’ve found corruption to be a way bigger issue on Windows. I had been using a Win10 install for about 5 years and eventually it just stopped booting and I had to reformat. Maybe it was my SSD, but I’ve been running Linux on that same SSD ever since then with 0 issues.
Realistically you don’t have to if you’re not constantly tinkering, but if you’re changing a lot of low-level stuff without knowing what you’re doing, you have the ability to break things. If you don’t know how to fix them, then it’s easier to just reformat. Basically it’s a skill issue lol.
I’ve broke things often and had to reinstall a lot because I didn’t know what I was doing. Still kinda don’t know, but do y’all recommend anyways to learn the knowledge?
Like I could probably read through man pages but I want something that shows how everything builds on each other to fill any gaps I’m missing
Depends on what you’re breaking I guess. If it’s DE stuff, kernel stuff, etc. Usually I just find a good YouTube tutorial if I want to learn something new and don’t know what I’m doing.
The Arch Linux Wiki is an incredible resource, even if you’re running another distro. Most of it is pretty universal (other than specific commands like the package manager), and it explains how everything functions and fits together. If I’m troubleshooting, it’s always my first stop.
I would recommend reading the manuals yes. Their are many manuals and not all are equal. The man pages can feel a bit strange as they list everything the software can do. To learn I found the archwiki to be better. (Also info manuals but many people are weirded out by the controls used to read these.)
Also don’t blame yourself for reinstalling if you mess up. It’s normal especially if you need the computer to actually work in a timely fashion
Just keep breaking stuff! It means your learning and trying new things, for the most part. Eventually you’ll just break stuff less and less or know what to look for when something breaks. On that note do try to struggle with something a little bit before rolling back or reinstalling.
Yep tinkering with the system is probably the main issue (for that NixOS is awesome btw.). But even when you’re not constantly tinkering. System-State accumulates over time, bugs are also apparent in (upgrading of) distros, and the maintainers of a distro cannot realistically handle every upgrade time-point x -> y, so stuff will likely break after some time.
But even when I have fixed all the issues in my previous at some time broken distros, at some point it just feels good to have a freshly installed system without all that dirty accumulated state (NixOS + impermanence and you’ll have that every reboot :P, see also this)
Yup. It’s a very manual install that’ll let you screw it up, so it’s gained that reputation. But it really isn’t bad if you follow the wiki (or have done it before).
That’s only if you use an automated script, and only if it works. ‘Default’ install is almost entirely manual, other than letting pacman grab what it needs to.
Arch has been daily driver for years, I’m already familiar with the process. There’s an option for a guided system. The default is a terminal with no guidance.
It depends. I use k8s a lot, and annoyingly enough we still happen to use cronjobs and bash scripts to “automate” certain tasks. Maybe it’s inertia, but bash is certainly easy to fall back to…
Yeah I am deep in the kube world as well. Since this industry-wide shift started happening, I feel like I write essentially no code anymore outside of bash scripts to glue things together. It’s essential but it’s not a replacement.
This cartoon seemed to me to be suggesting that you could implement the behavior of kube with bash. That’s obviously absurd.
Totally possible to go overboard in either spectrum of complexity. But yeah, take Prometheus for example. It’s super easy to set up and does a great job of metrics. Reimplementing this in bash would require… a lot of work.
Nope, I can run it on old potato with 3GB of RAM and i doubt i could run Ubuntu’s full snaps flotila . They also remove the telemetry of Ubuntu. But AFAIK you can turn on snaps. The way i understand it Mint has these main goals : get rid of questionable Ubuntu things, keep it super stable, be welcoming to newcommers (like my none tech parents who never seen Linux could just use Mint outbox the box)
For some reason, this didn’t work on my old phone after installing PixelExperience 11 on it.
There’s a third way. Bluetooth. At least you don’t need a cable, and you’ll save power.
For that reason, I usually use Bluetooth instead of Wi-Fi, unless I need higher bandwidth (except during peak hours of network usage, when my connection speed is below 1Mbps anyway).
I love it, because you can also get best out of both worlds in relation to the comic discusses. You can personalize OS to your liking, and the entire configuration is in a file, so you can redeploy the same setup again.
I feel like I keep posting this everywhere but there’s a project called AstOS that attempts this. Also someone clued me in on this distro neutral solution. AshOS. Full disclosure I haven’t used either.
It looks like solutions like these miss the whole point of what Nix is trying to do. Nix comes with the belief: “Unix has some fundamental issues, because it was designed in specific way. If we store things differently it works really well, and we even get those cool properties for free”.
The authors of those projects instead of thinking “this looks interesting, and it is a paradigm shift but it might be worth to to try feel like Linux noob for some time and start thinking a bit differently how the file system is structured to see if this change is really worth it”
Instead it is: “I don’t need to be PhD in Computer Science (whatever that means), here is how I can force this Nix feature or two on traditional Linux, with ansible, bubble gum and some duct tape and make it immutable-ish, which fails sometimes but, hey, it has the same feature on paper.”
Well to be fair I think it’s because they aren’t trying to be NixOS. You could leverage those arguments against any distro that’s trying out an immutable flavor. Which is mostly accomplished through btrfs features.
I agree that Nix/NixOS does a lot more and it’s a genuinely impressive and paradigm shifting project but it does break with traditional Linux layouts and thinking in a way that immutability doesn’t necessarily have to do.
You could also make the same argument with the systemd and non-systemd crowd.
Either way I look forward to the future of both immutability projects and NixOS. I feel like both areas still need a bit of work but they’re both really exciting fields.
Just add rescue to kernel options (if you use GRUB, press e to edit it for the current boot) and it will boot into console from which you can do downgrade.
When I started using Arch I just set it up on a btrfs filesystem and wrote a simple btrbk hook to take a snapshot before any package updates. That made it trivial to unfuck anything that broke after an update. I can’t remember the last time I had to roll the system back but it’s nice for peace of mind.
Start by playing with subvolumes and snapshots so you can get a feel for how they work. Once you’ve got that down you can break down your root filesystem into sensible subvolume chunks (/, /home, /var/log, /var/cache etc) so that you only snapshot relevant content during each update. I wrote a btrbk config at that point, tested it a few times and then wrote a pacman hook to fire it on install, update or package remove events and went from there.
Here’s what I use to take snapshots - you’ll need to write an appropriate btrbk config file for your subvolume layout but it’s otherwise feature complete. gitlab.com/arglebargle-arch/btrbk-autosnap
Like I mentioned above, I haven’t actually needed to roll the system back in ages but I get a lot of mileage out of being able to reach back in time and grab old versions of files for comparison.
Time shift is a lot easier if you’re just starting out but it also requires a specific subvolume structure and isn’t very flexible.
Edit: pro tip: don’t make /var a separate subvolume from /, it’s way, way, way too easy to roll one or the other (/ or /var) back without the other. If you do that by accident pacman’s state becomes out of sync with the running system and everything breaks. Stick to splitting frequently rewritten data like /var/cache and /var/log off, leave /var itself in the root subvolume.
Well yeah obviously like NixOS. My reason for not using it is that they use a non standard Linux filesystem and it renders a # of packages I want to install incompatible.
linuxmemes
Active
This magazine is from a federated server and may be incomplete. Browse more on the original instance.