Tmux saved me

Just wanted to spread the word of appreciation for tmux. I'm doing a big backup of our company's MinIO data. And we've currently undergoing a DDoS attack, so the connection isn't exactly great, ssh connection drops etc.
But I've started the backup session inside of a tmux, so when I eventually drop out I can just get back in with the help of `tmux attach`.
So, thank you all people pertaining to this piece of technology! I know there are other terminal multiplexers, namely screen, so this thanks goes to all of them! I'd recommend anybody who works over terminal to take a look into it, it's pretty easy to learn.

https://redd.it/1kbg3mc
@r_linux

I miss it...
https://redd.it/1kbekp5
@r_linux

Bytedance Proposes Faster Linux Inter-Process Communication With "Run Process As Library"
https://www.phoronix.com/news/Bytedance-Faster-Linux-IPC-RPAL

https://redd.it/1kbfls1
@r_linux

Ubuntu 25.04 & Fedora 42 Hit A Long Sought Milestone With HDR Support Working Well On The Linux Desktop
https://www.phoronix.com/review/linux-hdr-2025

https://redd.it/1kbjorz
@r_linux

Libreboot 25.04 "Corny Calamity" released! (free and open source BIOS/UEFI firmware replacement based on coreboot)
https://libreboot.org/news/libreboot2504.html

https://redd.it/1kbnfd7
@r_linux

I don't know why, but Ubuntu is looking crisp!

I updated/upgraded the packages today and notice a Wayland update. I don't know why, but man, the system is looking CRISP. Floorp/Firefox fonts just got so much better.

Anybody notice something like this? I am missing something or is just a thing of my "mind"? Also, is running more smooth than ever. (I deleted old kernels as well, maybe this improved the performance)

https://redd.it/1kbrw7m
@r_linux

Would you say that the Steam Deck is the biggest/most effective advertisement to encourage using Linux?
https://redd.it/1kbvdh6
@r_linux

Linux US market share at nearly 5%~
https://redd.it/1kc074t
@r_linux

Should I join the ride and switch to Linux?

Sorry, I'm one of those guys that came here from the PewDiePie video, what he put up is too good to be true, it looks magnificent, smooth, and full customization which I can't help but envy for, the managing of tabs and workspaces, it looks sick, it's great!

But I don't know if I should switch, and if the things I do will be affected, first off is I'm using an office laptop that can run quite well, it can play Dark Souls III no problem and high graphics Dota 2. I often do stream, but I usually just stream visual novels or Souls like and Dota 2, and maybe even drawing streams, can I still do that in Linux?

I'm thinking Linux mint, which is referred as beginner friendly, which is also what Mutahar recommended from SomeOrdinaryGamers. I also do office work, but just a basic Microsoft Word and Microsoft PowerPoint alternative would be sufficient for my office work, and of course the websites, I'm assuming I could also still open Microsoft teams through website right?

I've been using Windows 10 the whole time, and refuse Windows 11 because I tried it and hated it, and apparently my Windows 10 dies or stops getting maintenance on Oct this year, or something... I failed computer science class but for gaming and a sick software I'll gladly relearn it.

But if it's a no, I'll gladly be rejected if it would heavily affect what I do negatively, but if it's as good as what it's hyped up to be, with obviously some drawbacks, then I hope I'm welcomed.

And I also play Arknights on Bluestacks, and I refuse to play it on mobile.

can I play Elden Ring?

Edit: for drawing stuff I only use FireAlpaca and I don't have any license for a lot of things and just pirated some.... is utorrent available on linux?

To be fair my main goal is a slick smooth workspace that I don't have to wait minutes to open, but, if possible, I would still like to play my games, and possibly elden ring which it's very slow on my windows and crashes literally unplayable, but if that's impossible then it's fine.

I'm willing to spend days tinkering just to see my workspace looking smooth and slick asf, while still being playable. I also forgot to mention I play so much Terraria Modded as well so I'd be nice if there's also a performance enhancement, but I'm asking too much.

I haven't really seen Linux since back in like 2010 on computer classes that I didn't listen, didn't expect pewds to cover it and enlighten me with all these cool things.

I already saved all my important files on a separate hard drive and have an extra USB

https://redd.it/1kc283e
@r_linux

26 lines of Bash to edit notes with server syncing and encryption
https://github.com/megahomyak/micronotes

https://redd.it/1kc0h3j
@r_linux

Alternative Desktop Metaphor - Gnome

Out of all the popular desktop environments, Gnome is the only one that pushes for a modernized and innovative experience, ditching the traditional windows-like desktop.
At the same time, it is perhaps the most controversial DE; people either hate it or love it.
Do you think Gnome deserves its hate? If so, why, and do you think we need to innovate the traditional desktop worflow?
I personally think Gnome is at least decent.

https://redd.it/1kc1lw5
@r_linux

Are Linux distros converging?

I recently moved from Aurora Linux (based on fedora atomic) to Debian 13. My setup is nearly identical:
- kde plasma 6
- Kodi and other apps as flatpaks
- server apps as containers (Podman)
- cli apps as brew packages
- uv for python
- nvm for node
- firewall management via firewalld (pre installed)
- service management via systemd

I also have a MacBook and I use brew and oci containers in that machine.

Edit: and topgrade to update all my stuff

https://redd.it/1kc798n
@r_linux

I'm really liking the Ghostty terminal.

I feel over the past few years, terminals have become less customizable. In Gnome, transparency is a hidden pref! You get lots of predefined themes, but they're difficult to modify.

Recently, I wanted to rice my fastfetch output and I found only one terminal that accurately displays an image - **Ghostty**.

It's also easy to customize with just a dozen lines in a config file. (pasted below).

Anyway, if you miss being able to fine-tune the look of your terminal, give Ghosttty a try.

# Save to ~/.config/ghostty/config

window-height = "29"
window-width = "110"
quick-terminal-position = "center"
background = 000000
foreground = ffffff
background-opacity = 0.85
background-blur = true
font-family = "Intel One Mono Regular"
font-size = 14
window-padding-x = 9
cursor-style = "underline"
bold-is-bright = "true"



https://preview.redd.it/197xl9mf56ye1.png?width=2560&format=png&auto=webp&s=261ab2cafdd04110bd39c2289af64eb008f00865





https://redd.it/1kc8bn4
@r_linux

Redis is Open Source again
https://antirez.com/news/151

https://redd.it/1kcdc2j
@r_linux

TMUX user here... what's all the hype about latest Terminal Emulators?

I've been a TMUX user for years, and nothing will ever make me change. No, I don't need tabs...pshaw. I see all this hype for Ghostty, Wezterm, Kitty, etc. Is there any real advantage to using these over the default terminal that comes with the desktop I'm on? I feel like I might be missing something.

https://redd.it/1kccwv5
@r_linux

systemd-analyze blame doesn't say what you think it does

In my experience the systemd-analyze blame output is grossly misinterpreted all over the internet and it's influencing people to kneecap their systems in a misguided pursuit of efficiency.

OK, so let's say I'd like to improve the boot time of my system. Let's take a look:

$ systemd-analyze
Startup finished in 6.321s (firmware) + 529ms (loader) + 1.043s (kernel) + 3.566s (initrd) + 32.429s (userspace) = 43.891s
graphical.target reached after 32.429s in userspace.

32 seconds doesn't seem very good. Let's look at the blame output to find out the cause:

$ systemd-analyze blame | head -n5
30.021s lazy.service
4.117s sys-devices-pci0000:00-0000:00:1a.0-0000:05:00.0-nvme-nvme1-nvme1n1.device
4.117s dev-disk-by\x2dpath-pci\x2d0000:05:00.0\x2dnvme\x2d1.device
4.117s dev-disk-by\x2did-nvme\x2dnvme.1987\x2d3436394630373138314537303030303034393739\x2d53616272656e7420526f636b657420342e3020325442\x2d00000001.device
4.117s dev-nvme1n1.device

Oof, 30 seconds!? That has to be it! Let's see:

$ systemctl cat lazy.service
# /etc/systemd/system/lazy.service
Unit
Denoscription=a very slow service

Service
Type=oneshot
ExecStart=/usr/bin/sleep 30
RemainAfterExit=yes

Install
WantedBy=multi-user.target

$ journalctl -b --no-hostname -o short-precise -u lazy.service
May 01 08:39:31.852947 systemd1: Starting a very slow service...
May 01 08:40:01.874683 systemd1: Finished a very slow service.

Yep that takes 30 seconds alright. But is it making my "boot" time slow? What happens when I reboot? After logging in I'll check systemctl status:

$ systemctl status | head -n5
...
State: starting
Units: 347 loaded (incl. loaded aliases)
Jobs: 3 queued
Failed: 0 units

We're still starting up as I write this reddit post — lazy.service has not yet finished! That's because the userspace time reported by systemd-analyze and the startup time reported by blame don't correspond to the "boot" time at all by colloquial usage of the word: I could have logged in, started firefox, checked my email, and written this whole post before my system "booted". Instead, blame is reporting on all the tasks that systemd executes in parallel at startup time, including those that can continue to run in the background.

Crucially, many services' (e.g. udev-settle, wait-online, etc.) only explicit purpose is to wait and watch for some event to occur so that subsequent services can be started. For example, Time and time again users notice that something like systemd-networkd-wait-online.service appears near the top of the blame output and go about disabling it. This service uses event polling to be notified when a network connection is available, so that subsequently started services are more likely to complete a successful connection immediately instead of after several attempts. An alternative strategy like exponential backoff implemented as a fallback in most networked applications is much slower because you are waiting during the time when the network becomes available practically by definition. Technically you could disable this service, but that makes your observable "startup time", the time before your startup applications start doing useful work, quicker, not slower. The numbers don't matter.

Something like systemd-analyze critical-chain systemd-user-sessions could be helpful, but it has several caveats as noted in the manpage, in particular that it only tracks start jobs for units that have an "activating" state. For example, the following output:

$ systemd-analyze critical-chain initrd-switch-root.target
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the

systemd-analyze blame doesn't say what you think it does

In my experience the `systemd-analyze blame` output is grossly misinterpreted all over the internet and it's influencing people to kneecap their systems in a misguided pursuit of efficiency.

OK, so let's say I'd like to improve the boot time of my system. Let's take a look:

$ systemd-analyze
Startup finished in 6.321s (firmware) + 529ms (loader) + 1.043s (kernel) + 3.566s (initrd) + 32.429s (userspace) = 43.891s
graphical.target reached after 32.429s in userspace.

32 seconds doesn't seem very good. Let's look at the blame output to find out the cause:

$ systemd-analyze blame | head -n5
30.021s lazy.service
4.117s sys-devices-pci0000:00-0000:00:1a.0-0000:05:00.0-nvme-nvme1-nvme1n1.device
4.117s dev-disk-by\x2dpath-pci\x2d0000:05:00.0\x2dnvme\x2d1.device
4.117s dev-disk-by\x2did-nvme\x2dnvme.1987\x2d3436394630373138314537303030303034393739\x2d53616272656e7420526f636b657420342e3020325442\x2d00000001.device
4.117s dev-nvme1n1.device

Oof, 30 seconds!? That has to be it! Let's see:

$ systemctl cat lazy.service
# /etc/systemd/system/lazy.service
[Unit]
Denoscription=a very slow service

[Service]
Type=oneshot
ExecStart=/usr/bin/sleep 30
RemainAfterExit=yes

[Install]
WantedBy=multi-user.target

$ journalctl -b --no-hostname -o short-precise -u lazy.service
May 01 08:39:31.852947 systemd[1]: Starting a very slow service...
May 01 08:40:01.874683 systemd[1]: Finished a very slow service.

Yep that takes 30 seconds alright. But is it making my "boot" time slow? What happens when I reboot? After logging in I'll check `systemctl status`:

$ systemctl status | head -n5
[...]
State: starting
Units: 347 loaded (incl. loaded aliases)
Jobs: 3 queued
Failed: 0 units

We're still starting up as I write this reddit post — lazy.service has not yet finished! That's because the userspace time reported by systemd-analyze and the startup time reported by blame don't correspond to the "boot" time at all by colloquial usage of the word: I could have logged in, started firefox, checked my email, and written this whole post before my system "booted". Instead, blame is reporting on _all_ the tasks that systemd executes in parallel at startup time, including those that can continue to run in the background.

Crucially, many services' (e.g. udev-settle, wait-online, etc.) _only_ explicit purpose is to wait and watch for some event to occur so that subsequent services can be started. For example, [Time](https://www.reddit.com/r/archlinux/comments/1hclnjf/systemdnetworkdwaitonlineservice_takes_much/m1s19yp/) and [time](https://www.reddit.com/r/linux/comments/1kber3v/so_i_noticed_many_dont_know_about_the/) again users notice that something like `systemd-networkd-wait-online.service` appears near the top of the blame output and go about disabling it. This service uses event polling to be notified when a network connection is available, so that subsequently started services are more likely to complete a successful connection _immediately_ instead of after several attempts. An alternative strategy like exponential backoff implemented as a fallback in most networked applications is much slower because you are waiting during the time when the network becomes available practically by definition. Technically you could disable this service, but that makes your observable "startup time", the time before your startup applications start doing useful work, _quicker_, not slower. The numbers don't matter.

Something like `systemd-analyze critical-chain systemd-user-sessions` could be helpful, but it has several caveats as noted in the manpage, in particular that it only tracks start jobs for units that have an "activating" state. For example, the following output:

$ systemd-analyze critical-chain initrd-switch-root.target
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the

"+" character.

initrd-switch-root.target
└─systemd-tmpfiles-setup.service @2.290s +54ms
└─systemd-journal-flush.service @1.312s +957ms
└─var-log.mount @1.302s +7ms
└─local-fs-pre.target @371ms
[...]
└─system.slice
└─-.slice

shows the startup time of some units in the initrd, but completely misses that the bulk of time in the initrd was waiting for amdgpu to initialize, since its a udevd stop job that waits on this action:

$ journalctl -b --no-hostname _RUNTIME_SCOPE=initrd _KERNEL_DEVICE=+pci:0000:03:00.0 -o short-delta
[ 1.162480 ] kernel: pci 0000:03:00.0: [1002:73df] type 00 class 0x030000 PCIe Legacy Endpoint
[...]
[ 1.163978 < 0.000039 >] kernel: pci 0000:03:00.0: vgaarb: VGA device added: decodes=io+mem,owns=none,locks=none
[ 2.714032 < 1.550054 >] kernel: amdgpu 0000:03:00.0: enabling device (0006 -> 0007)
[ 4.430921 < 1.716889 >] kernel: amdgpu 0000:03:00.0: amdgpu: detected ip block number 0 <nv_common>
$ journalctl -b --no-hostname _RUNTIME_SCOPE=initrd -u systemd-udevd -o short-delta
[ 1.160106 ] systemd-udevd[279]: Using default interface naming scheme 'v257'.
[ 2.981538 < 1.821432 >] systemd[1]: Stopping Rule-based Manager for Device Events and Files...
[ 4.442122 < 1.460584 >] systemd[1]: systemd-udevd.service: Deactivated successfully.
[ 4.442276 < 0.000154 >] systemd[1]: Stopped Rule-based Manager for Device Events and Files.
[ 4.442382 < 0.000106 >] systemd[1]: systemd-udevd.service: Consumed 3.242s CPU time, 24.7M memory peak.

So eliminating these services would not be faster. These commands are useful, but just make sure you actually have a problem before trying to fix it.

https://redd.it/1kcg7b0
@r_linux

Sriracha - Imageboard and forum
https://codeberg.org/tslocum/sriracha

https://redd.it/1kcdata
@r_linux

Got stuck in vi

So I have a story to tell. My phone suddenly decided to get stuck infinitely booting after usual force shutdown because of low battery. While it was failing to boot, adb was accessible. I decided to try adb shell in it. Android apparently has vi preinstalled, because of course it does. And while I was realizing I couldn't access su, I SOMEHOW typed vi and hit Enter. In one moment I realized what all these "Can't exit vim" memes were about. In one moment my ego shattered. Ctrl+C didn't work, Ctrl+Q didn't work, Ctrl+X didn't work. I was stuck. I literally had to google how to exit it. I haven't been more betrayed by Android than that.

#Step bro I'm stuck, help me exit vi

https://redd.it/1kcjrym
@r_linux

Flathub: A paradigm shift for distributing applications — Jordan Petridis at LAS 2025
https://inv.nadeko.net/watch?v=NxOH4wJkfLY

https://redd.it/1kcgnps
@r_linux