First release candidate of systemd 258 is here Like it or not, systemd is the industry-standard init system these days. A new release is coming, and it's a big one. Version 257 of systemd came out last December, and since the team tends to release new versions twice a year, you could interpret that as meaning that version 258 is slightly overdue. The reason could be that …
Works here too, but I'm not a power user by any means.
The trouble is that if I need to search the net for how to do something, increasingly the answer involves systemd commands, most commonly to restart a service.
Most of the internet seems to ignore Devuan, even ventoy won't run it's installation ISO image (last time I tried). Looks like we're going to be left out in the cold.
I have posted about Alpine here a few times - it was up against Void and I got it working first on a Pi5, but I havent given up on Void. That is another minimal distro, but is a bit more "alien" in terms of package manager and init system. I'll be trying it soon on an old Acer laptop.
This. So much. I guess we should start contributing to the Devuan forum, at least that would be a central place. I now try to document my problems and solution attempts locally, so I can retrace my steps when messing up because there is no documented fix for what I'm trying to do. Though Devuan mostly does what I need.
I have been burned by "Agent P's" software several times (pulseaudio... that was a steaming pile of mammoth droppings when it was forced upon me by some distro, seems more stable now, and the resolvd and networking portion, as well as some temp file stuff of systemd), more than by anything in my neigh on three decades of Linux use (except maybe sendmail? The configuration was a nightmare for me the first times), and while I get that the init system should be redesigned I don't think putting every functionality in it is the best way....
Most of the internet seems to ignore Devuan ...
Indeed ...
Like it has been noted in more than one opportunity, here and elsewhere: there is a fantastic amount of moolah behind the making of systemd the de-facto init for the Linux ecosystem.
The likes of MS/IBM/RedHat have put all their weight behind it so it comes a no surprise when all search engines respond did you mean Debian? when you search for Devuan.
Soon they won't even ask ...
But the hoi polloi does not give a monkey's toss about all that and it makes sense that this is so.
Why should it be different?
With their mugs stuffed into a smartphone screen all day they get all they (think) they need.
Because, yes, it is so very convenient !
Systemd is nothing short of an MS registry for Linux, its main purpose being to turn Linux into a MS type OS, with Debian eventually being totally absorbed by it.
But there's really nothing new here: it is the old MS embrace, extend, and extinguish that has been going on for decades.
Only that now there's active participation and backing from IBM/RH, with all that it entails and cooperation from the inside, so to speak.
Devuan (and derivatives) is still holding on but who knows for how long this will be so.
Meanwhile, in Linux land, very few people remember what the principles of Unix philosophy are and stand on.
.
More like... it's installation image doesn't work with Ventoy, not the other way around. Ventoy is just a boot loader. I ran into one that didn't work, it was because the distro boot image was trying to mount its /sources directory internally and it was trying to mount one of the ventoy partitions instead. I had to write the image directly to a USB stick (I used dd).
That's not a ventoy problem, it's because of the way the Devuan ISO is formatted. I think the live installer works.
The desktop live ISO works for me on Ventoy and my IODD...the netinstall ISO is a mixed bag though.
The trouble with these contrarian distros is that eventually they end up looking so far back they stray towards the line of not supported by anything.
The concept of these distros is fine, but in practice, they usually end up built around badly maintained forks of older tech, not the actual older tech itself...because they have to keep patching and fixing the old tech to ensure it still works with newer stuff. The folks maintaining these distro specific forks have neither the skill nor talent that the OG developers of the dead project possessed and you end up with something that is much worse than the original tech.
the problem (in SystemD) is FINDING WHAT SERVICE NAME, amongst the bowels of THE BLOB [SystemD] - is not possible to just grep the init.d files after having done a ps with The BLOB, now is it...? and so you do MORE searching than necessary because it is Micros~1 style monolithic BLOATWARE.
Anyone who likes SystemD, try figuring out how to turn off everything that interferes with a serial port that you are trying to use to command an external device... I'll wait. [the list is pretty long AND not obvious - had to research this on an RPi a few years ago and it seemed to take forever - bloody patch on the wall where I kept slamming my head in FRUSTRATION]
the 'service' command also works in FreeBSD
This post has been deleted by its author
Devuan has a bug, which the team feels is not their problem, but only happens in their implementation.
On mate desktop using mate-media, the devuan implementation does not persist your audio device settings choice across reboots.
The problem does not occur if you install antix, or convert debian to devuan, but devuan think it is a but in mate??!
I personally cannot use devuan because of this bug.
Have you submitted a PR? The community is very supportive.
Yes, I know this is "learn to code" for distros. But honestly, "oh it doesn't persist my settings" sounds like a very fixable problem. At the very least, submitting it as a bug and then helping to reproduce it will almost certainly result in happiness.
I've been using Devuan since very early--the only real problem with the whole project is that they picked a name that autocorrects to Debian.
Devuan is generally the only Linux distro I use these days [except for embedded things where I may have little choice]
It's not infected by "The Blob" - my new name for SystemD)
If this new login-within-a-login "multi home dir" crap gets in the way of application development, I'll just REFUSE TO SUPPORT IT!
"Is there any truth to the rumor that Systemd release 259 will include a spreadsheet and word processor?"
Those two applications are typically already baseline crap so probably no.
Next up is incorporating one of Microsoft's LLMs into systemd to second guess your system calls and shared library bindings amongst other unsolicited AI assistance. :)
"Gee, just as I was deciding to switch to Linux..."
Gee indeed. Same here: Windows user strongly "encouraged" by Windows 11 (and 12 and 13...) to finally and seriously prepare for Making The Jump.
Had more or less decided on Mint+Cinnamon; being aware of, but so far graciously overlooking, the systemd thing.
Now this article and the comments generate a strong expectation of "out of the frying pan into the fire".
But what to do if these features are desired:
-- semi-Windows-like or MintCinnamonish L&F (but not MS-like T&C :-))
-- no high-brow hackery required (at least initially, for installation and setup)
-- Ventoy supported (for the transition phase)
-- and no systemd
Devuan a bit on its way out (as per this discussion)?
Alpine a bit too minimalistic (ditto)?
MX a bit of a dark horse (for me)?
Any other suggestions for a Plan B?
What says the Elevated Hive Mind on El Reg?
Cheers.
"Devuan a bit on its way out (as per this discussion)?"
No. As ususal its tracking Debian so whenever Trixie becomes the current Debian Excalibur will follow it as Devuan current a few weeks later as usual and it's OK to use now, in fact. The usual bits of enshittification in UI due to devs everywhere not being able to leave well alone - one of them seems to be KDE playing games with default window opening behaviour. I suppose they thought it was time the swindow system menu has some more work to do. Still, it could be a lot, lot worse. It could be Windows.
Yes SystemD is worse than init in many ways. Yes it's trying to take over the job of the OS
But at least with SystemD you still have a level of ultimate control over it via config files etc - unlike with Windows. So it's still better than Windows IMO.
For day-to-day use you'll be perfectly happy with Mint + Cinnamon I'm sure. Try out the other distros by all means (my recommendation would always be to spin them up in a VM first to see how you find them before committing to wiping anything) but you'll likely find that you don't get much benefit from avoiding SystemD.
I'm sorry but why the ^^^^ would you want multiple Home folders for a single account? This is made especially pointless since you have to choose which one you want at login which means you could instead use 'child' accounts (Fred1, Fred2 for example) to get the same effect without needing to rewrite the core code.
In general this list of changes seems to be a traditional mix of WTF? ; single scenario value; someone's pet idea they have been trying to include for years and finally a sprinkling of probably useful but may already be possible in way everyone already understands
"it's a lot of little commands that have well defined purposes that you are not required to use if your use case does not require you to run them"
There's one small difference between that at Unix. With Unix they were separate commands, not bits sticking out of a hairball.
They're separate commands in systemd. This is quite obvious simply by looking at the source code. They might reside in a single repo. It doesn't mean there is dependencies between them and quite frankly it is why most of the raging about it is so comically wrong.
I love how many whingers leap out of the woodwork whenever wayland, systemd or rust are mentioned on this site. Lurking & waiting to pounce with the downvotes when they're corrected on their nonsense as if it means anything or validates their ridiculous opinions. Meanwhile Linux marches on oblivious to their whining.
Back in the day, they also worked – those little commands.
Need I mention my last run-in with systemd which ... er... broke `sudo`. Yes. It stopped `sudo` from working so I couldn't even fix it without booting from a live image, chroot, and then replacing the entire thing with Gentoo without systemd.
And, the WHOLE time, I was asking myself this question: WHY does anything have any reason, means or motivation to even be able to break `sudo` in the first place and, as a corollary, why should *I* need to know about its very existence?
Sure – there's a tonne of those old-back-in-the-day UNIX commands that I don't know and don't use – mostly, ones I've likely even forgotten I ever did know or did use because my memory's like that – but, if you don't call them, they do nothing. They do not break things that aren't in their realm and there is no requirement to learn about some facet of them because – surprise! – someone else made an executive decision that it should now break something you've always used, known, and relied upon.
(And, yes, it was `systemd-homed` that broke `sudo`!)
I got half way through this article and realised that I actually do not care to read about all the wide, sweeping, revolutionary changes. I do *not* want wide, sweeping, revolutionary changes on my Linux boxen and that's why I will not tolerate `systemd` on them, either. I just wish that more people in the Linux world would look at those massive piles of stuff that it's bringing to the table and ask: who wants that? Instead of 99,99% just going "systemd is everything" and swallowing it.
If anything ends the momentum of the Linux desktop, it isn't going to be nVIDIA's drivers, Wayland, Steam failing to bring gaming to Penguin-land, Microsoft indulging in anti-competitive malarkey or Linus pissing off one too many developer or contributor: it's going to be a complete collapse of the entire thing because nothing independent from systemd even exists any more and so the house of extremely inflammable cards not only grows but becomes more and more critical in the system. Any risk manager who's not jumping up and down about that should just buy macs for their server farm and office desks!
"WHY does anything have any reason, means or motivation to even be able to break `sudo` in the first place"
That's an easy one: because it breaks Salzer and Shroeder's Separation of privilege principle: "Where feasible, a protection mechanism that requires two keys to unlock it is more robust and flexible than one that allows access to the presenter of only a single key." https://en.wikipedia.org/wiki/The_Protection_of_Information_in_Computer_Systems
I doubt they made such an improvement deliberately and in any case the whole thing breaks the Economy of mechanism principle: Keep the design as simple and small as possible.
> I do *not* want wide, sweeping, revolutionary changes on my Linux boxen
Then you should look at the BSDs.
Linux is a corporate toy now, an enterprise tool used in billions of phones and servers. It happens to run on desktops too -- I'm typing on it -- and it's not bad at it but it's rapidly getting enshittified along with all other mass-produced toys.
The biggest thing to happen to Windows in a decade is embracing and extending Linux.
If you want time-tested tools built the old fashioned way then this is not the OS for you.
If you want a good modern 64-bit OS for modern 64-bit kit, FreeBSD.
If you want proper old style stuff, even more minimal, then NetBSD.
If you want security and cleanliness above all else and are happy to give up fripperies like Bluetooth to do it, if you're the sort of person (like me) who favours wired peripherals anyway, then OpenBSD.
And it tries to do everything as one giant monolithic app. Hmm, where have I seen that design philosophy before?
It's classic overengineering: oh someday somebody might want to use a Sharpie to take some notes while editing magic, byzantine config files, so it needs to have integrated support for ordering Sharpies from AWS, and not as a pluggable add-on but built in to the kernel with every install. What if they want colored sharpies? We need to add support for that. What about generic brands? In fact, let's just add general support for online shopping of all items, not just writing tools, but let's make sure there are sharpie-specific options in the first release that are later obsoleted and cause it to quietly fail without logging an explanation why.
But I digress. Does SystemD have a built-in email client yet?
I'm completely agnostic. I want whatever just works that is regularly maintained by a team of people that care about it. I mean, X11 was a bit of a "kitchen sink" situation...and yet it's still the darling of contrarians everywhere. There is massive overlap with the people that hate systemd and the people that hate Wayland...these same people will complain about feature creep in SystemD but not in X11...folks don't know what the fuck they're talking about half the time.
The mask slipped with the Xorg team recently when someone tried to fork X11. You can Google the story yourself...but what it boils down to is someone came along that was prolific with providing bug fixes, updates etc etc to the X11 codebase after may years of declining commits etc basically breathed a bit of life into X11...some of the existing team members started getting pissed off and salty at this guys insane pace. Some regressions crept in, sure, but the sheer volume of positive improvements far outweighed the occasional bug.
Projects die and go out of favour for many reasons...they aren't always technical reasons...sometimes projects are forked or replaced because of the people managing the incumbent project not necessarily because the tech sucks. Personally, I think the team behind a project are as important as the project itself...if the team gets a bit too big for it's boots and starts hindering progress, then regardless of the tech, it's time to move on.
I don't have a problem with systemd at all...it works and for me, it's no more complicated to use / manage than previous init systems...there was a bit of a learning curve to begin with, but these days I just get on with it and it has never got in my way. I don't love it, it's a fucking init system...why would anyone "love" an init system? I don't hate it either...it is what it is and it's fine for what it does.
For me, it's more important that the team behind a project is functioning correctly and is open to outsiders coming in and getting shit done, patches rolled out, bugs fixed etc. The minute a team becomes a bit gatekeeper-ish like the X11 guys...it's game over because that tells me that they no longer give a fuck about the project itself and they're more interested in keeping a club going.
As for people complaining about having to learn new things and port shit on a semi-regular basis...this is fucking Linux...I've personally never stopped learning for 20+ years...it's part and parcel of being on Linux.
>> why the ^^^^ would you want multiple Home folders for a single account
To make it more like Windows, where the user directories are more and .org obfuscated, and take directories like Documents or Pictures can be anywhere, and that files that look to be local are really remote (see OneDrive), and when you lose your internet connection, you lose access to your files.
I'll go back to FreeBSD before using systemd.
This is needed when you want all people to have access to the same resources and you want to track their login/activity as separate activites.
Yes, this could all be handled by using chmod, groups, etc. but if you're from a Windows world (poor bastards) this is how they work.
I've didn't mind MS-DOS 3.x, but it all went downhill after that. Unix was always a better solution for everything but gaming. And now with Steam that's been fixed I see no reason to touch the OS that is the demon spawn of Redmond!
Actually Windows-95 wasn't all that bad -- After one installed about two dozen "Service Packs" (Service Pack=Massive collection of bug fixes). It was actually comprehensible to almost all users and could run for weeks without crashing. Of course it was really MSDOS 6.22 with a GUI shell. We have since devolved into the incomprehensible shambles that is Windows-11
> Windows-95 wasn't all that bad -- After one installed about two dozen "Service Packs"
Your memory deceives you.
Windows 95 never ever got a single service pack. Neither did Office 95.
Nor did 98.
It got new versions. Most cost money. Most couldn't be upgraded to without hacks; you had to reinstall.
95... 95A (with FAT32)... 95B (with USB)... 95C (with IE4)....
98 got 98SE but it cost money.
Office 97 is when it started shipping so full of bugs it needed Service Releases. This is also the version that had the first ever release of Outlook.
I submit these are not coincidences.
I agree with your overall verdict, though: for the time, W95 was a work of art.
But no, it wasn't DOS 6.22 with a wrapper. It was DOS 7.00 and the OS was a lot more than a shell. The GUI was the real OS. DOS was the filesystem layer during bootup.
95B got DOS 7.1. That was the end.
I bought a PC in early 1998 that came with- what a bit of checking confirms was- the final version of Windows 95, OEM Service Release (OSR) 2.5. From my experience, it was very close to Windows 98 in all but name, certainly far closer to that than it was to the original release of 95.
From memory- which Wikipedia backs up- what I remember being odd was that (as their name would suggest) these later versions of Windows 95 were *only* available as OEM versions installed on a new computer. If you went out and paid full whack for the retail-packaged version of Windows 95- even then- you got the three-year-old original.
I've no idea why that was.
Much as I hate systemd, and all my computers run systemd-free Linux or Unix, there *might* be a use for this feature, if each home folder has its own passphrase which can be disabled, deleted, set, etc. by system admin(s).
In a corporate environment, you desire a 1:1 mapping between users (human) and user accounts. This way, if someone is suspended or fired, Accounts/Security has just ONE account they need to find and disable.
Simultaneously, it happens that a single person may have multiple roles, or jobs within the company, each role or job having a separate set of valid filesystem accesses.
Further, there are things which happen in login scripts which may be required for Department A, yet should not happen for Department B.
The current kluge-workarounds are:
(1) multiple accounts, e.x., jonesf1, jonesf2, etc , which violate the desired person:account 1:1 mapping;
(2) file permissions granted via groups. The problems there are you still have the login script conflict, there are not enough groups, and group access is insufficiently fine-grained;
(3) AFAIK, not possible under Unix and Linux: file access permissions granted to specific users, as is possible under Novell Directory Services, and possibly (IDK) under Microsoft Active Directory. This is an administrative nightmare when a person is suspended from their job in Department A, but still works for Department B.
Systemd areanames might help resolve the issue.
One of the problems of systemd is that you are required to use and maintain the entire 218-square-mile Cape Canaveral launch complex, even if you're only shooting off a bottle-rocket.
> it happens that a single person may have multiple roles, or jobs within the company, each role or job having a separate set of valid filesystem accesses.
Which is what groups gives you.
> there are not enough groups, and group access is insufficiently fine-grained
This means you haven't bothered to analyse your requirements and defined, then created and used, enough groups to control your own specific environment. Ok, there are limitations from some file systems (NFS, IIRC) but kludgey careful use of groups gets around those (i.e. define the set of groups important to whatever is being accessed via (NFS or ...) to be within the range you can use there).
Nor is it clear how having a different home folder helps - aside from selecting which login script gets run and which ~/.myapp-config.d gets invoked, which are separate from shared resource access control, aren't they?
> Further, there are things which happen in login scripts which may be required for Department A, yet should not happen for Department B.
Can you elaborate on that, as it is not something I've encountered in practice (simply because I've not worked in every environment on the planet, not because it is necessarily unusual): do you mean w.r.t. job function - e.g. logging into the correct time management system? - or physical constraints - e.g. Dept A all use printer PrintA, Dept B uses PrintB?
Either way, it really isn't clear to me why adding this into the OS ((making it more complex for everyone) really helps in such situations instead of, say, shoving into their login scripts a question box that just asks "What are doing today" then does the necessary.
(3) AFAIK, not possible under Unix and Linux: file access permissions granted to specific users, as is possible under Novell Directory Services, and possibly (IDK) under Microsoft Active Directory. This is an administrative nightmare when a person is suspended from their job in Department A, but still works for Department B.
Of course you can , facls.
The world went to Hell in a hand basket when people started calling directories folders (For the young:- Directories point to where things can be found in a file system, folders contain bits of paper).
I suppose systemd might make sense to us if we were a few decades younger and didn't know cleaner ways of doing things.
Not really a problem. Your storage is really a patchwork of blocks of data. The file system is an abstraction layer to present that as a coherent structure of directories and files (with links added to give a bit of fleximility. A GUI interface is another layer of abastraction layer on top of that and folders are part of that. You think "abstraction layers" are just fancy talk? If so, ask yourself if you really want to deal with your computer as the mass of ones and noughts it really is because everyithing that makes it easier and less confusing than that is an apstraction layer and they're piled one on top of another from ragiststers counters and ISAs upwards.
While I don't disagree with you, since the lowest common denominator of the designers of abstraction layers passed the point of "files and directories are a too abstract set of ideas for our target user, let's assume they can only cope with something more concrete that they're likely to understand from daily experience," the world did go to hell in a handbasket.
I wish him well with his pottyOS operating system, but Linux it ain't!
The biggest annoyance is it's considered a Linux upgrade path rather than a different OS.
Why does it remind me of "this new software you need is out, but you must upgrade to windows to use it".
Develop pottyOS to your heart's content, but don't force it on Linux users. It's classic embrace, extend, extinguish.
"In post #22, Agent P explains that he uses this a lot and so has added a parameter to expand the new VM's filesystems. The new switch --grow-image= (or -G for short) does it for you, so you don't need to manually adjust the size with truncate or fallocate. This is something we've never once felt the urge to do, but perhaps it will be good news for someone somewhere."
Growing the rootfs partition & filesystem is standard practice for Linux "cloud" disk images, you have a OS disk image sized just as necessary for the actual files it contains, then when you deploy variously sized VMs using that disk image the rootfs is automatically grown upon 1st boot to fill the rest of each VM's storage device, typically via cloud-init's growpart/resizefs modules.
This is not cloud-specific, you can do the same with local VMs.
So whilst "This is something we've never once felt the urge to do" if you've deployed Linux VMs on AWS, Azure, GCP, etc it is fairly likely that this auto-resizing has occurred.
Knowing this, and wishing to build a systemd-free system (because it seems to have crept up on me unawares as I progressed from Mandrake to Mageia via Mandriva), I built a new system for my business and used PCLinuxOS instead.
Generally speaking it works fine on a very modern GigaByte mobo and I don't miss systemd at all. The only "issue" (so far) that seems to follow on from losing systemd is that I can't use ExpressVPN on it, plus there are a small number of apps that are on my main Mageia machine but not on PCLinuxOS, but both are running Plasma 5 and there are few other real differences once I have tweaked it to look the same.
There have been a few crashes, but I have been able to handle that (and having run Linux since 2005, you would certainly think that I should by now). The main function of the system is to support voice recording and it seems to do this quite well, with Focusrite support now in the kernel and not therefore needing further software to enable it.
I did check out a LiveDVD of Devuan previously and was quite impressed, but having run rpm-based systems for so long, I wanted to stay in that fold, hence PCLinuxOS rather than just another installation of Mageia. But there seems to be little trouble living without systemd.
There has been one utter drag about PCLinuxOS (apart from their server building apparently going up in flames recently), and that is that the fresh installation defaults to Wayland and logs me out of my desktop each time I do Ctrl-C (and I absolutely live on Ctrl-C), so I have to either uninstall kwin-wayland or otherwise hobble it by tweaking different files after installing. Otherwise, it runs fine.
Maybe I shouldn't have mentioned W*yla*nd in the same breaths as systemd...
The core function of systemd is to act as a service manager... which is why it is factually inaccurate to describe it as an "init system".
I've been using Linux exclusively on my own machines, and then exclusively-exclusively when I retired, for almost 25 years. And I have never had an essential system process that crashed and needed to be restarted. OK, I'm just a home user, but none of my machines need a "service manager".
Although a long-time Debian and eventually Devuan user, I've started trying out Alpine on low-spec machines. From doing "ps -ax" I see that it seems to have a "service manager". I'll bet that its codebase is a tiny fraction of systemd's.
The next release might as well contain a kernel and call itself an OS.
So much for 'do one thing and do it well': this continual scope creep of systemd should have us all on edge. It's the modern twist on the time-honoured dependency issue xkcd covered so very well.
At least Midgely was solving some real problems and his solutions did genuinely work; it was only afterwards that we realised there was potential for problems with one of them (the lead was probably foreseeable). So Midgely wasn't ever facing a stream of vocal criticism each time his work was released.
But if Agent P. is blissfully unaware of criticism...
"Agent P. is blissfully unaware of criticism..."
He IS totally aware of the criticism ... BUT DOES NOT CARE !!!
This is a person who KNOWS he is right and therefore any comments that do not agree with his own opinions are simply ignored !!
He is RIGHT and you are WRONG, because you are either stupid or willfully being obstructive !!!
I am constantly surprised that all these HUGE egos [Putin, Trump, Musk 'Agent P' etc etc etc] can co-exist on this small planet !!!
:)
I like this snippet from his wiki entry:
Midgley's legacy is tied in with the negative environmental impact of leaded gasoline and freon. Environmental historian J. R. McNeill opined that Midgley "had more adverse impact on the atmosphere than any other single organism in Earth's history", and Bill Bryson remarked that Midgley possessed "an instinct for the regrettable that was almost uncanny". Fred Pearce, writing for New Scientist, described Midgley as a "one-man environmental disaster".
There certainly are striking similarities to systemd though that's probably a bit harsh on Midgley.
"had more adverse impact on the atmosphere than any other single organism in Earth's history"
There's an interesting point here. The biggest impact any organism had was the first photosynthetic one. It and its decendents released an extremly reactive molecule into the environment that totally changed the environment of the planet. Whether this was adverse or not could be debated as an abstract point: we have a vested interest in it but all the obligate anaerobes descended from pre-photosynthetic life have been driven into restricted niches.
My appetite for this arcanery† was sated by #10.
Even the names are becoming silly.
Systemd-confext just looks like a typo or leaves one wondering how one, apart from Father Jack, could fext a system ?
Systemd.covfefe ?
† or arguably chicanery.
> Systemd-confext just looks like a typo or leaves one wondering how one, apart from Father Jack, could fext a system ?
“systemd-sysext activates/deactivates system extension images. System extension images may – dynamically at runtime — extend the /usr/ and /opt/ directory hierarchies with additional files.”
I didn't know that SystemD was now a file system /s
It's quite clear that systemd is adding functions for two reasons:
One: P. wants it on his laptop
Two: His employers want it for their cloud
No one cares about anything else. Like things that work, things that are not bloated, philosophy, etc.
The more tangled shit Agent P builds into his emissions, the more Debian developers will twitch. More, more, the sooner the better, time to waken the dead-heads.
"We are normal and we dig Bert Weedon. We are normal and we want our freedom." -- Bonzo Dog Doo-Dah Band, a very long time ago.
Before having to use a distro with systemd, I'd read the horror stories and prepared myself for the worst. I was quite surprised how well it installed, and ran without any real issue. Until it didn't. Then it became the shitshow I had been promised. Debugging a startup component was a nightmare, fixing it even worse. It's like when you move into a new house and your neighbour seems like a fine bloke. Come round for a beer, borrow my lawnmower. Then he rapes your wife and murders your children.
<sings>
Young man
You feel down on your luck
I said young man
you're mad at Linux
Young Man
Why can't they see
You don't wanna use systemd
I said young man
You don't have to give up
Ya don't need Windows,
Mac or Linux
Young man
You can be free
Of RedHat forcing systemd
<bah bah bah>
Freee B-S-D
I said
Free B-S-D
It doesn't use systemd
It's quite Unixy
It'll make you shout out with glee
Freeeee B-S-D
Y'know I'm saying
Free B-S-D
It's not like MS at all
Systemd don't make the calls
It's fun and works for you and me....
I'll see myself out, thanks
I'd forgotten that he's an MS employee. It all makes more sense once that connection is established. It's a shame good (bad) old Bill Gates has moved on. I'll never forget his rant about Movie Maker (and in passing, Windows Registry, which is what systemd mostly reminds me of)
https://www.techemails.com/p/bill-gates-tries-to-install-movie-maker
One failure in SystemD can potentially bring down multiple components, such as services, logging, and networking, creates a single point of failure that undermines reliability.
--
+ log_warning("[color="Red"]/usr appears to be on a different file system than /. This is not supported anymore.[/color] "
+ "Some things will probably break (sometimes even silently) in mysterious ways.");
The comment section of this article was exactly what I expected to be: overwhelmingly attacks on systemd and Lennart Poettering, coupled with pointers to Devuan. Raw meat. Sharks.
I must admit, I'm not a professional sysadmin. I manage a large number of my own Linux devices, and they work well for me. If systemd helps make that happen (and I do use its facilities quite regularly), that's fine. I wouldn't design a system the way that systemd is designed, because I do agree with many of the design criticisms of systemd, but it is usable. If all of the people who object to systemd were to get together and design an init system they liked, they might come up with something good. Maybe it would be better not just than OpenRC and Upstart, but systemd as well. If that were the case, then distros would have at least some incentive to replace systemd with this new thing.
I might mention that Linux is full of cruft, as any 30-year-old software system would be, so why not produce a Linux distro without systemd AND Linux? That would definitely be cruft-free.
I think you did a pretty good job attacking systemd, honestly.
Toss in the fact that Poettering's ego is overbloated or the notion that his entire career has consisted of him badly ripping off Apple designs, and you'd have a pretty typical systemd comment yourself.
Rather than recommending Devuan, I recommend The UNIX-HATERS Handbook. Reading heretical texts is a pretty solid way to break oneself out of a dogmatic worldview.
I'll join in on receiving some of the hate, but have an upthing votermawhatsit anyway.
"f all of the people who object to systemd were to get together and design an init system they liked, they might come up with something good"
That's not necessary. It already exists. It's Syvinit. Systemd is exactly the sort of thing you get when you lose sight of "good enough".
Or OpenRC which I run on my development workstations and all my servers, without a single problem. Literally: I have *never* had a real problem that stemmed from the init system because that's literally all it is: an init system.
Sure: it sometimes inits something that is broken because sometimes other things don't work but that's not inits fault.
The reason systemd is a problem is purely because it is purportedly there to be one thing but, in reality, does a whole lot of other things that no significant audience is wanting. You can't mobilise an army of init experts to "solve systemd" because, at best, that army will produce the best possible init system which will entirely fail to gain traction because being the best init system isn't even the same game or competition. Similarly, there is no army of experts aiming to build the best kitchen sink to compete with systemd on its own terms because no experienced experts believe that a kitchen sink is good or even viable.
So, when industry comes along, they funnel funding, convention, norms, backing and popular vote into their own pet kitchen sink and systemd is what comes out the drain.
The Bell Labs/BSD init is even better: none of that unbelievably stupid run-level shit.
sysvinit was when the lose sight of "good enough" started and the trend of fixing stuff that wasn't broken took off.. And it's been downhill ever since.
I found the responses above hilarious. There is a better init system than systemd, it's OpenRC, no it's sysvinit, no it's the original Bell Labs init. Not one of the responses was about requirements., and clearly there isn't agreement on which of those is best. I might mention that back in the days when I used a PDP-8, it didn't have an init system at all. Or an operating system at all, come to think of it.
To respond to a couple of other points. I don't indulge in personal attacks on people, and have no opinion one way or the other about Lennart Poettering. I am sure he's a perfectly fine person, though, and I take at face value his desire to write useful system software. I might also mention that there are many others credited with work on systemd, so if there is a conspiracy, it's a big one. Maybe it involves tunnels under a Washington DC pizzeria? The github project contributors page goes on and on with different people who have had PRs accepted. Somewhere I saw a figure that there have been something like 2000 or more different contributors.
As for a claim that I was dumping on systemd, I wasn't. I do use it, and rarely get frustrated with it. Yes, it isn't done the way I would have done it, but there are loads of Unix things I wouldn't have done the way their authors did them: man pages written in nroff markup, for example. My criterion for excellence in software is not `is it what I would have done?' but `is it useful?'. And I find systemd eminently useful.
Now that I'm old and grey, I just use operating systems, rather than write bits of them. I am a Linux user and have been with Debian almost since there was Debian. It's got Systemd. It works just fine and causes me no headaches, ever. My Debian server is only rebooted for kernel updates - and not very often those - and more frequently, power outages. It works incredibly well and reliably. There are more important things to get heated about than OS init systems. BSD is fine too, but I have over 25 years of Linux learning and can't be bothered to start again. I had a quick look, out of interest, and couldn't really see the point. Flame away, see if I care.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
