Diseased Open Source Software Community - it's about ethics in Code of Conducts

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
resulted in corrupted data on each unit restart
This isn't systemd's fault, this is your own shitty software. Learn to write crash-only software, noob. Not that systemd's abhorent behavior in this case is justified, but "Our java process received a sudden SIGKILL" is the most common thing in the universe because of the OOMKiller.

My problems with systemd are multifaceted; It's in far too many things and it's far too many pieces put together. It's your mounttab, it's your cron, it's what controls user login. Do these pieces have complex dependencies? Yes. Does that mean you should give up on having abstraction layers and bundle them into one monolith with a web of tendrils throughout> Fuck no.

Give me an init system and a mounttab that support layer tagging, so that they can ping-pong back and forth, rather that one where sshd is dynamically injected with dbus bullcrap so it can receive "events" or whatever.
 
It works on my machine too, but that's because I did not obsess over it autistically. Maybe if you read everything I said you would understand this.
Maybe I will say it differently. There is only one Windows PC in my house atm, and it's work laptop.
And even this Laptop is only used as a Citrix launcher to do actual work on Linux machine somewhere else.
Linux is perfectly usable nowadays, there are very few things that Windows or MacOS has over it nowadays, as for most people OS is just base for web browser anyway.

I don't remember encountering issue in my daily use of Linux for a long time. It works fine, it does what it has to do.
Moreover I had more issues with Windows in recent year than with it. Just today Edge decided that it will ignore audio inputs for good know why reason.
Unless you are grabbed by the balls by Adobe, I really don't understand that sentiment that "Linux is not good OS for doing actual work", what's wrong with it again?
That it can come barebones? That was just your autistic choice.
 
I really don't understand that sentiment that "Linux is not good OS for doing actual work"
I should have clarified that I'm talking about Arch-based distros since they're so autistic.

My general frustration with computers (software specifically) has pushed me over the (autistic) edge and now I just want to make my own OS.
I am aware of how ridiculous and contradictory this sounds. "So, you don't like barebones systems so you will make your own system?"
What frustrates me is that with nigger systems I am given almost no control over what happens, but with linux I am not given enough.
Basically, if I write the OS I have maximum control and can tweak everything to my needs.

In retrospect the above sounds super retarded.
 
for most people OS is just base for web browser anyway.
Legitimately I think most people would be better off with ChromeOS. It's funny to me that people are totally fine spending $1,000 on a phone, but don't want to spend more than the absolute minimum on a laptop (desktops are for fuckin' nerds btw, what is this? the 90s?). It's staggering to me that people actually buy those crappy laptops that have like 4GB of ram, 64GB of shitty eMMC memory, a Pentium, and Windows S mode. Legitimately, they'd be way better off with a Chromebook with the same specs.
Windows is a Chromium bootloader for the average retard, that is when they have to do serious "work" like using Office 365 or Microsoft Teams (also available on iOS :)). The average normie of course abhors actually using their computer for internet usage. That's what their phone is for! Duh!
 
There is no practical reason to install everything in the system yourself, installing arch takes 20 minutes at most, but what's the point? You learn nothing by doing so, there is no practical benefit, and it's just a waste of time. (Not to mention Gentoo, but at least their wiki is decent.) And then, after everything is installed, you just get a barebones system and have to install the WM and other programs yourself. (Also fuck pulseaudio, fucking nigger trash.)
I should have clarified that I'm talking about Arch-based distros since they're so autistic.

My general frustration with computers (software specifically) has pushed me over the (autistic) edge and now I just want to make my own OS.
I am aware of how ridiculous and contradictory this sounds. "So, you don't like barebones systems so you will make your own system?"
What frustrates me is that with nigger systems I am given almost no control over what happens, but with linux I am not given enough.
Basically, if I write the OS I have maximum control and can tweak everything to my needs.

In retrospect the above sounds super retarded.
You don't want to waste any time installing Gentoo, now you want to build your own OS.

You are a special kind of retard.
 
Linux is perfectly usable nowadays,
If you use it like you would a basic windows/apple laptop, sure. Youtube, basic coding, email, whatever, you can do that on windows, macos, linux, who cares.
The moment, you try anything else, you are... well, i wouldn't say screwed but you'd see why so many people prefer windows.
I fucked up my linux install twice, and by fucked up i mean, i had to boot to the installer to get myself a CLI to fix my collossal fuck up.,
Both of them involving screwing over the graphics, once wehen i messed around with Display managers, the other when i messed around with my nvidia X server settings . That was in the last year.
What's the last time you collosally fucked up your windows install simply by changing the graphics settings around, to the point where you had to reinstall it.
Let's not even get started on the comparativly "minor" issues, from wlan not working as it should, to audacity not recording nor playing sounds (a issue i just finished fixing, took me 2 and a half hours, and i consider this a minor, "day to day" issue). Then we have gaming which is a roulette, especially if you want your shitty nvidia mobile 1050 to work, which in some cases it completely refuses. I almost shat my kernel up trying to install nvidia drivers. (doesn't help that my broke balkan ass has to pirate games)
So yeah, linux is perfectly usable if you want to use it as a chromebook./raspberryPI/server.
You get used to it eventually, but tbe moment i started messing with windows 10 VM's i promptly got reminded why people prefer windows. Because windows just, works. If you ignore all the AI bullshit, it works, so does a macbook, no bullshit necessary, whether it be gaming, general use, whatever. It. Simply. Works. And that right there is what normal people want. A OS that simply just works. Why? Because macOS and Windows are made by companies with more money than first world economies, but linux is made by a large group of deranged autists who can't agree on shit and constantly sbotage eachother.
 
You don't want to waste any time installing Gentoo per your first message, now you want to build your own OS.

You are a special kind of retard.
I addressed your criticism already.
Installing Gentoo wasn't a waste of time the first time I did it because the wiki and install guide teaches you some things about the system, unlike the Arch wiki. But I will never install it nor will I use it again. The emerge times are simply not worth the extra performance.
 
Windows is a Chromium bootloader for the average retard, that is when they have to do serious "work" like using Office 365 or Microsoft Teams (also available on iOS :)). The average normie of course abhors actually using their computer for internet usage. That's what their phone is for! Duh!
I just got a laptop to do work for a new customer. It's Windows. But the Office 365 apps are only the cloud versions, accessed with a web browser. Apparently the department has to pay extra to IT if they want me to have 'actual' Microsoft products. Since I'm doing development work it doesn't really matter anyway.
 
I addressed your criticism already.
Installing Gentoo wasn't a waste of time the first time I did it because the wiki and install guide teaches you some things about the system, unlike the Arch wiki. But I will never install it nor will I use it again. The emerge times are simply not worth the extra performance.
And what would be building your packages for your custom OS? NASA? AMD going to drop you an EPYC server to throw in a datacenter to become a build server? Whose going to manage package upgrades, patches, version compatibility?

So an emerge takes too long. Yet you want to build an OS from scratch (which requires compiling everything from source), yet that will be okay.

Gentoo has binary packages and has had them for a year plus now.
 
  • Like
Reactions: UERISIMILITUDO
This isn't systemd's fault, this is your own shitty software.
BULL-FUCKING-SHIT. Go run mariadb or postgres and start sending SIGKILL to them midtransaction and come back telling me they're badly designed. Or better yet, yank the powercord since that would be the equivalent of how systemd treats children processes. Why the fuck would you design it this way? What's the point of sending SIGTERM to the main process and SIGKILL *in the same fucking moment* to children?

Not to mention i had to design what today would be called a docker entrypoint, because who needs env variable expansions, RIGHT?
 
BULL-FUCKING-SHIT. Go run mariadb or postgres and start sending SIGKILL to them midtransaction and come back telling me they're badly designed. Or better yet, yank the powercord since that would be the equivalent of how systemd treats children processes. Not to mention i had to design it this way because who needs env variable expansions, RIGHT?
Unrelated to systemd, but either of those should be able to handle getting randomly killed.
 
Unrelated to systemd, but either of those should be able to handle getting randomly killed.
Older versions of mysql (pre-innodb) would absolutely shit itself. Newer ones will probably recover depending on which exact op it would be. Postgres will complain and possibly rollback the wal transaction. Neither of the above should be expected to be caused by a fucking init system.

Better yet, disable all journaling and async on XFS, start writing (complaining how it's now dog slow) to disk then pull the plug. Once you realize your filesystem is trashed go ahead and mail Ted Tso how his filesystem is shit because it should obviously be powerloss resistant without all the mechanisms that make it so 🎩
 
Newer ones will probably recover depending on which exact op it would be.
Does it not have a WAL? Its not exactly difficult to make it atomic.
Neither of the above should be expected to be caused by a fucking init system.
I don't really want to touch the systemd argument because I don't do anything overly complex with it (so its fine for my usage), but SIGKILL with no timeout seems like the wrong answer.
 
Unrelated to systemd, but either of those should be able to handle getting randomly killed.
And they do. I have literally pulled the power plug on a running production postgres server to demonstrate failover, pointed out the brief spike of errors and latency we saw in alerts, and pointed out how brief it was

100% your software's problem. Go read about Crash Only Software. Go read Postgres' documentation on reliability, which explicitly calls that out.

One aspect of reliable operation is that all data recorded by a committed transaction should be stored in a nonvolatile area that is safe from power loss, operating system failure, and hardware failure (except failure of the nonvolatile area itself, of course). Successfully writing the data to the computer's permanent storage (disk drive or equivalent) ordinarily meets this requirement. In fact, even if a computer is fatally damaged, if the disk drives survive they can be moved to another computer with similar hardware and all committed transactions will remain intact.

If we follow this procedure, we do not need to flush data pages to disk on every transaction commit, because we know that in the event of a crash we will be able to recover the database using the log: any changes that have not been applied to the data pages can be redone from the WAL records. (This is roll-forward recovery, also known as REDO.)
 
Does it not have a WAL? Its not exactly difficult to make it atomic.
It does, but there are still cornercases in innodb
but SIGKILL with no timeout seems like the wrong answer.
It's not that it's all SIGKILL, let me give you a concrete example. Here's how it goes down

Consider a s6-svc daemon. This is a bogstandard init-lite system that spawns underlying processes, let's say you have some python workers accepting data on a http socket and writing that to a file. So the process graph in the cgroup would look something like this
Code:
s6-svc (500)
|- python (501)
|- python (502)
|- python (503)

Now, you send a systemctl stop to that. s6 gets a SIGTERM, so it'll start cleaning up processes 501-503. But lo and behold they are gone! Why? Because systemd killed them the moment it got the stop command. Now your files are fucked, s6 is confused, but not to worry because it'll get a SIGKILL soon enough. But hey, your stop command worked! No zombie processes!
What should really happen is PID 500 should get a SIGTERM, which will then be propagated to 501-503 and ONLY IF that fails after a timeout, send a SIGKILL to *all*of them.

100% your software's problem.
There's a difference with this occuring on accident and happening by design. I hope you also tested on couple-hundred-gigabytes worths of datasets that have to be scanned on every "crash". Thankfully the people in charge agree with me.
 
Last edited:
Consider a s6-svc daemon. This is a bogstandard init-lite system that spawns underlying processes, let's say you have some python workers accepting data on a http socket and writing that to a file. So the process graph in the cgroup would look something like this
Code:
s6-svc (500)
|- python (501)
|- python (502)
|- python (503)

Now, you send a systemctl stop to that. s6 gets a SIGTERM, so it'll start cleaning up processes 501-503. But lo and behold they are gone! Why? Because systemd killed them the moment it got the stop command. Now your files are fucked, s6 is confused, but not to worry because it'll get a SIGKILL soon enough. But hey, your stop command worked! No zombie processes!
What should really happen is PID 500 should get a SIGTERM, which will then be propagated to 501-503 and ONLY IF that fails after a timeout, send a SIGKILL to *all*of them.
I can't think of a good reason for it to do that. That's retarded.
 
Just because something CAN survive an init system doing something stupid doesn't mean it SHOULD have to do it as a regular occurrence.

My systemd experiences are usually the opposite... "Shutdown" .... "Waiting for something.... 1 minute of 3 minute timeout" Why, there was literally nothing running except a couple terminals.
 
Back