You must log in or register to comment.
Imagine having backups and not being on the testing branch of the beta version of a distro while running a custom kernel that is on alpha (Context, im on testing branch of fedora 39 beta with the asahi kernel)
Everybody in here does all this crazy shit with their system. I just wanna use my computer, man. I cruise on defaults all day long. I barely even bother changing the DE’s default wallpaper.
Tl;dr:
If you do tweak your system debian seems the most stable one.
Ok I switched to full Linux no Windows about one and a half year ago.
First I tried an Ubuntu gaming variant. It wasn’t working like I wanted and outdated. Then manjaro because it was said to be good for gaming and easier than arch. I couldn’t get warm with it too many hurdles to get stuff going. Fedora or rather nobara (from the same guy who makes glorious eggroll for Proton) was my choice then I really liked it and it worked mostly like I wanted. But because it is basically dependend on RedHat and they went closed source and I had issues (which weren’t solved by a new distro, I messed up my kde configs) I switched to debian-testing.
I knew debian well because it’s the same I run for years on my old Laptop which wouldn’t Support Windows 10.
And I must say Debian-testing is great, stable and up to date with drivers and stuff. I had to do a few steps to get steam running and install flatpak but then it’s just the best experience I ever had on Linux.
What I actually wanted to say is that I usually do a bit of tweaking and then break sth. But on debian I didn’t need to do that and if I did it still works fine.
How is it compared to ubuntu?
As long as you don’t reach too far up the OS’s ass, Debian should be more stable compared to Ubuntu and its derivatives just because it isn’t as preloaded with stuff you might not need.
Besides, Canonical is just another Red Hat waiting to happen.
Colourful metaphor but accurate. Try to do things The Debian Way first and you’ll rarely get into trouble. Start screwing with existing packages and you break assumptions fast.
I thought we were all in agreement that it had already happened
Well I use Fedora but you are probably fine with most distros. There’s Linux mint if you just want everything to work.
Same. I’m not looking at the wallpaper anyways, I’m staring at software all day long instead. It just can’t be too bright otherwise I flashbang myself at night.
Am I too NixOS to understand?
Or ZFS/btrfs snapshots. But yes, NixOS does it right.
Timeshift can make use of BTRFS snapshots btw
btrfs + snapper + pacman-hook = happiness
I just use btrfs and snapper.
I “use” it too. I don’t really do anything, it came setup like this out of the box with my distro and it just does it’s thing until I mess up hard.
I swear I used timeshift so much when using a NVIDIA card
Ditto.
- Install OS, timeshift.
- Get Nvidia, CUDA and CuDnn all playing together, timeshift.
- Install everything else safe in the knowledge that no matter how badly I break things at least I won’t have to do step 2 again.
Or just like, you know, learn about what you did n fix it .-.
What’s a good overview of time shift?
It makes a copy of your entire system automatically (and your home folder if you want it to) so, in the event that you break something and can’t/don’t want to fix it, you can go back to your most recent back-up from before you messed your system up. I’ve had to use it a few times because I installed some drivers for my drawing tablet that broke more than they fixed and I didn’t want to deal with the pain in the ass of removing them and all of the dependencies they installed.
How is this different than a regular backup? Not salty, just curious.
I believe it’s using a feature built-in directly in the filesystem.
I’m just curious if it’s possible to browse individual snapshots like in MacOS Time Machine and fetch individual files out.
Now mind you, everything I write might be wrong, I am out of my depth here.
But as I understand a BTRFS snapshot is simply a (subvolume in which you will find) copy of the table that points to the actual files or, rather, blocks on your drive. As long as a table exists that points to a block, this block will persist.
The nature of BTRFS is Copy-on-Write so in your active snapshot, when you modify a file / block, a copy of it is created with the new version, referencing this new block on the filesystem table.
This is why BTRFS snapshots are fast and take little space by themselves, you do not need to actually copy all the data at the moment of creating the backup, rather when the data is modified and only that data.
Timeshift saves!
When you use a journaled filesystem … what are backups?
Oh the number of times that I have tweaked something only to only have to start over is too damn high!
Timeshift worked exactly once for me, and by once I mean it messsd up my entire system so I had to install something else instead
If you don’t test your restores, you don’t actually have a backup. You have to test the config first to make sure it works for how your system is setup. An all defaults system should work out of the box, but if you start to alter and customize your system in ways that the backup is not configured to handle, you are in deep risk.
i hate tinkering so i would never do anything that disturbs the system itself, timeahift advertises itself as a backup system, never seen anywhere that said backup needs to be tested, whatever that even means. besides i shouldn’t have to do that to begin with.
never seen anywhere that said backup needs to be tested
I don’t know what to tell you, you must be really new, because recovery tests are the first thing that is said when discussing backups. It’s the backbone of systems operations. If you don’t test your backups for recoverability, you really don’t have backups at all. is a widespread saying in the tech industry.
Daily timeshift ftw!
When you have backup kernel:
Is that a problem?
