Adventures of a Linux Noob

I had a bunch of old nix generations I wasn't using, cleaned those up and got hundeeds of gigabytes of free space

I've been in a similar situation

edit: For context, there was a bug with the graphics driver that was putting out an error every frame, at 200+ fps... needless to say, I could actively see the log growing in size

Holy shit
- right? what the hell?
This is what confuses me about Linux defaults, why would it let them grow that large?
We can tune logging settings to resonable values for the max size and everything, it just doesn't come that way for some reason.
- If you don't use archaic technologies it actually does. systemd-journald is limited to max(10% FS size, 4GB) per default.
- Well, Linux is also made for servers and super computers, and just imagine it refusing to keep logs because the file's too large
Mmmm somebody need some log rotate in their life.
Oh my production s***'s on point but all the Dave and QA s*** I need at least one failure before I get around to doing law rotate. I guess I should spend the time to make the ansible job.

93 GB is like one weekend of moderate media piracy for me...

If it makes you fell any better, after doing a fresh install, I tried a "finally finished setting everything up" backup and was immediately out of space.

Turns out it was saving backups to my boot sector. 🤦🤦🤦🤦🤦🤦

we are all noobs in some regard. I've been using linux for private and work for 3 years and I don't know shit about tineshift. linux is such a diverse ecosystem and there's so many places to make mistakes and learn. It never stops. I fully expect to be bricking my machine on accident well into my 60s

Learning about new things is the best thing about Linux. I keep a folder with screenshots and saved html pages for all the fixes, workarounds and settings I've accumulated over the two years I've used Linux on my desktop. Highly recommend keeping a similar folder.
- Yup.
  Every time I fix something difficult I document it in great detail in Obsidian. It's a good feeling of, ''I'll never have to be confused by this problem again''.
  I reference it constantly too, so it isn't a waste of time. The waste of time would be not doing it.
- Everytime I stumble upon something it take some quick notes and put "I should start blogging this" on my bucket list. Then immediately forget about the blogging part until I take the next note...
- I keep a text file with all useful commands on the desktop and have a alias in terminal to access it quickly via nano. Works very well.
I have been using Linux for over 20 years and this post is the first time I've heard about timeshift. I use Arch, btw.
- ironically, arch users are the only users who I've heard talking about timeshift because apparently its the best way to roll back after an update breaks sth?

The humble 50GB /var/cache/pacman/pkg on my 256GB drive.

Every. Time.
Filelight: /var/cache/pacman/pkg 👀
Paccache helps, but sometimes a nuke is needed to clean up pkg.

400 gb of timeshift backups.. It felt so good removing them XDDD

Best thing that I’ve ever done was to write automate a weekly script that makes a ZFS snapshot and then deletes any that are over a month old.

That's a very good idea. Might wanna keep an additional yearly one too though, in case you don't use the computer actively for a while and realize you have to go back more than a month at some point.
- Ya, I offsite backup the entire zpool once a year at least. I have quarterly and yearly snapshots too.
  But the weeklies have saved me on several occasions, the others haven’t been needed yet.

pip cache is another common culprit, I've seen up to 50GB

node_modules has entered the chat
There was once a 220 GB log file on my pi-hole server. Probably was a bug though.

I use borg btw.

BTRFS + Snapper + BTRFS assistant has been pretty good for me

"Backups" to the same disk?

Personally i use them in case an update messes stuff up. So i can restore it back.
- So not backups, but snapshots.

lol, you can just set it up to keep the latest snapshots only.

(noob here)

Same like me who never realized i have so many BTRFS Snapper backup in every end of year.

I recently realized I forgot to use reflink copy on an XFS filesystem and ran duperemove which freed ~600GB of data

I think Timeshift has an option to only store n snapshots and to auto delete the oldest one if it's hit n snapshots. Or at least that's how my timeshift behaves but I set it up a while ago and don't remember the details of how

Yeah, it does have an option to keep x amount of snapshots. But the ones that took all the room I had created manually. They were from like two years ago and so I clearly didn't need them (Yes, I've been on the same Linux Mint installation for two years).
- Yes, I've been on the same Linux Mint installation for two years
  That's good! You shouldn't need to be reinstalling your OS all the time—a good operating system should be able to be installed once and the user not feel the need to reinstall or change it out for the duration of their hardware working. (2 years is also not that long anyway.)
  I suppose with your problem it's not that bad as you can just have a quick look as to what's taking up so much space and then delete accordingly. Perhaps Timeshift ought to automatically delete/rotate all snapshots except for ones the user explicitly says they want to keep, not just ones they made manually but might be fine with getting auto deleted.
Exactly. I have a daily, weekly and monthly backup that erases the previous version.

"dust" is my go-to cli thing for finding what's taking up hard drive space.

Speaking of, I should check my timeshirt settings

ncdu or gtfo

I recently had 2/3 of drive space taken by btrfs snapshots. Still learning to manage them properly :D

i am also running out of disk space. {pacman package cache, Team Fortress 2, a Windows VM and the android SDK being the main culprits.}

How small is your disk?
- 512GB.

Is there no way to have timeshift auto delete ? I guess thats not how it works huh

It does have autodelete, dont remember of its the default, but i think it explicitly offers you the option when you set it up

I usually set up a completely separate partition on a different drive for Timeshift. That way it doesn't gradually eat away at system space on the main drive. And even if it was on there, it would have already eaten all that space in readiness, so to speak.

Also, I don't have it backing up my home directory. I do that separately.

But that said, this post has given me the reminder to see if there are any old snapshots that could do with deleting. And there were a few. It's now back down to roughly the same size as my main OS install again, which is about as big as it needs to be if you think about it.

Yes, deleting all the old backup Linux kernels was something I had done before in a 128GB SSD. I then had to delete the Time shift entries as well to actually get that space back.

For me it's failed AUR comps