Systemd redux: The end of Linux

http://blog.lusis.org/blog/2014/11/20/systemd-redux/

by bcantrill • 12 years ago

235 444 12 years ago

blog.lusis.org

api• 12 years ago

One thing I've realized about the Linux community through all this systemd flame warring is how unbelievably conservative a large subsection of it is. There's this huge so-called "neckbeard" continent that views anything architecturally beyond the 1980s as a huge affront to Unix.

IMHO I kind of shrug at this, since Unix was never really all that great to begin with. Unix won because the only commercially viable and well supported alternative was Windows, an OS that was (and in many ways still is) significantly worse especially for server and embedded applications. Everyone rallied around Unix and especially free/open Unix as an alternative, and so here we are.

It's also tough to compete with free, and Unix OSes got a huge boost from both Linux and the various free flavors of BSD. Yet that boost came at the expense of things like BeOS, Plan9, original NeXT, and the OS I still feel is hiding behind the JVM ... which for their day represented fresh ideas that might have gone somewhere.

Ultimately I think the existing Unix paradigm is going to be killed by Docker and mobile OSes that containerize in similar ways, and I'm not sure this is a step forward. It escapes much of the ugliness and the poor permission model of Unix, but it does so by handing virtually everything to the app. Docker containers (and mobile apps) can be thought of as something almost akin to giant statically linked binaries. We're getting more monolithic and coarse-grained.

califield• 12 years ago

We're getting more monolithic and coarse-grained.

At the same time we are pushing more heterogenous software stacks to production and configuring more specific dependencies for our applications.

It almost seems like you're using cross-platform as a pejorative. ;)

baldfat7• 12 years ago

Yeah we need to go back to when distro A and B had totally different versions of everything since it was so much work to get things working.

Now we are close to having a OS where you can seriously just expect anything "Linux" to just run. Bad I guess to some :P

thasmin• 12 years ago

People who aren't extremely familiar with how the Linux init system works and whose job doesn't include keeping the servers stable don't see why the neckbeards are up in arms about systemd, but there's good reason. Many peoples' jobs depends on making sure the servers are working, and knowing how the servers work is a big part of their job (and sanity). In its current incarnation, systemd changes the fundamentals of how servers work without much increase in features - large risk and little reward.

johnny2210• 12 years ago

I deal with sysadmin stuff for quite some time(since 2001), and I just won't use distributions without systemd.

The benefits far outweigh the risks (imo obviously)

toolz4• 12 years ago

Server admins have so much new technology to learn and play with to stay relevant. If they feel like learning systemd is a chore, they might be in the wrong business.

zanny4• 12 years ago

I've had huge productivity gains with unit files for systemd over trying to write spaghetti shell code for old sysv. The syntax and features are well documented and writing them is extremely simple. I also don't believe any sysv implementation had crash recovery or socket activation of daemons, both of which are huge feature wins.

rodgerd5• 12 years ago

People who aren't extremely familiar with how the Linux init system works and whose job doesn't include keeping the servers stable don't see why the neckbeards are up in arms about systemd

Hi, I'm a sysadmin who's fed up with neckbeards (most of who apparently don't know much and refuse to learn) claiming to speak for all sysadmins on this topic.

large risk and little reward.

It's four years old, and claiming "large risk and little reward" is like listening to someone claim that moving from sendmail to postfix would be a disaster.

lucian19003• 12 years ago

Servers is precisely where I want systemd!

There are some things I've wanted reliable and consistent mechanisms for so long: starting/restarting/inspecting services, isolation/resource limiting, socket activation, log collection.

convivialdingo• 12 years ago

Well, here's the basic deal. If we're talking about common servers, common desktop, etc then systemd is an excellent replacement. It covers the base of users quite well.

But lets say you are building a highly specialized application. You are going to be making quite a few customizations which are far more manageable through a shell scripting environment than by customizing a bunch of binaries.

I assume that Redhat is going to cover a lot of the bases for most users out there. But for those of us in highly customized environments it's going to suck.

johnny2210• 12 years ago

that's just speculation until you clarify. Can you be more specific?

egil• 12 years ago

The status quo for such projects is to use Busybox, and I reckon it will continue for projects where systemd etc is too much.

derefr• 12 years ago

it escapes much of the ugliness and the poor permission model of Unix, but it does so by handing virtually everything to the app.

If you want to see the ultimate extent of this, look at the Wii and Wii U. Each game ships with an "IOS": effectively an OS kernel+initrd update package. Every game boots to the newest IOS available, so if one game updates to IOS v6653, then another game that only shipped with IOS v6652 will find the newer version on disk and use it.

However, a game's IOS requirement doesn't just have a version; it also has a slot. Each console has space for 256 individual copies of IOS, which are each independently versioned. So if two games both use IOS[58], then the game providing v6653 will overwrite v6652 on disk, and then the game providing v6652 will boot into v6653. But if one game is providing a version of IOS[58], and the other is providing a version of IOS[61], then their effects on one-another are isolated.

You can think of it a bit like the IOS codebase having 256 branches, and each piece of software being able to specify which branch of the kernel it was developed on. It gets the newest kernel released on that branch.

This allows a sort of "move fast and break things" approach to kernel development, where a kernel can be hacked to support new software in a way that breaks old software: you just stick your modified kernel into an as-yet-unused IOS slot, and old software will have nothing to worry about. This approach has resulted in my own (pretty unused) Wii U having ~73 different IOS slots populated with kernels.

Interestingly, if you think about it, this is pretty much a continuation of what Nintendo was allowing developers to do before: shipping random collections of chips in their own cartridges that DMA to the console, effectively creating their own extended console to run upon. Allowing your software to ship its own kernel is basically the software equivalent.

amaranth• 12 years ago

This is how the Wii works but the Wii U has a more traditional kernel model. This is required due to the system's multitasking abilities (run "mini" apps like the browser or download manager along side a normal app/game). The only time the Wii U uses the IOS model is when it boot into vWii mode to run Wii apps which also disables the new features of the Wii U software.

rodgerd5• 12 years ago

Unix won because the only commercially viable and well supported alternative was Windows,

Unix didn't beat Windows. Unix beat VMS and LISP machines and AS/400 and various other minicomputer operating systems. In fact, if we're talking about mainline commercial Unixes, NT started beating the shit out of them in the late 90s - if Unix lovers hadn't had the free ixen (Linux, BSDs) to fall back on it would be a sad state of affairs indeed.

arca_vorago• 12 years ago

AS/400

Hey now, I'll have you know AS/400 is still alive in going in my workplace! We also have an entire position just for it's programming...

ratboy666• 12 years ago

http://thorstenball.com/blog/2014/11/20/unicorn-unix-magic-t...

is also a current story on HN.

Just an example of how powerful that simple 70's Unix is. Allows features that appear "magical" to thorsten, anyway.

Windows? Wasn't really even an option... until 20 years later. And, of course, Unix really isn't that good, either. But, before you ignore it, please come to feature parity, at least.

zvrba• 12 years ago

Windows? Wasn't really even an option... until 20 years later.

Which 'Windows'? Before Windows 2000, there was Windows and Windows NT, the former being a more-or-less just a shell running on top of DOS.

teekert• 12 years ago

I completely agree. Having played with Docker on CoreOS for some weeks now I see that it will push a much bigger and different change than systemd, which, on my Arch box was just an update with no real problems, I just had to learn some new commands.

Docker though... man how different it is and how clean it makes my system feel. I do feel that Docker will move towards some kind of Docker optimized minimal dockers that are not Debian or Ubuntu or what ever, that is just a stage so you feel some familiarity.

CoreOS meanwhile, who will ever touch its init system except to auto-start containers? Which will be done by nice tool which hides systemd in the future I guess.

Ok, my posts does focus on the server side of course.

pjmlp5• 12 years ago

One thing I've realized about the Linux community through all this systemd flame warring is how unbelievably conservative a large subsection of it is. There's this huge so-called "neckbeard" continent that views anything architecturally beyond the 1980s as a huge affront to Unix.

Fully agree. It seems some people are quite happy with a few xterms in the X-Windows replicating a twm user experience, stuck in the past.

I would also add Oberon, Active Oberon, Singularity, Verve and the current unikernel/library OS research.

OS I still feel is hiding behind the JVM

Android kind of got us there. Now with Java being compiled to native code, maybe other C++ layers might be replaced in future versions, given how Android team looks at the NDK.

All in all, I want the Xerox PARC and Douglas Engelbart's visions, not the AT&T one.

keyle• 12 years ago

Can someone put in Layman's terms what the change entails and why it would be the end of Linux?

hvs• 12 years ago

I think he means, "the end of Linux as we know it", but that doesn't have the hyperbolic ring of the current title.

fidotron• 12 years ago

It's not the end of Linux. It's the final death of the idea that Linux should just idly continue to act like a clone of some vague Unix of old because it was better when men were men and their computers had obscure RISC processors with billions of registers.

Linux hasn't really been that way for an incredibly long time, but a large proportion of the userbase still cling to this notion, some for ideological reasons, and some out of sentimentality.

nitrogen11• 12 years ago

Systemd is most problematic because of its own-everything monolithic nature, and the changing of kernel interfaces to match. The vast majority of Linux systems are not desktops, and probably not servers either. The kernel and the core utilities around it (like init) should be designed with multiple implementations and all machine types in mind. This means phones, set top boxes, TVs, routers, automation hardware, supercomputers, and more.

If the core utilities around the kernel and the interfaces they use are well-documented, designed, and modular, then different machine types can pick and choose the components they need, and easily write their own by opening a well-defined API on something in /dev. If systemd continues to take over and alternatives smothered, very quickly the kernel will become useless to any system where systemd makes no sense, especially non-desktop embedded systems.

Aardwolf• 12 years ago

Why did systemd decide to ALSO take over logging and store them in a binary format, though?

blinkingled• 12 years ago

I don't agree with the article that it is 'end' of Linux, whatever that may mean, but the article argues that because SystemD is changing what Linux stood for - simplicity, never break userspace, keep as many things in user space as practical, and keep the kernel version dependencies on userspace to a bare minimum and in addition is also rewriting lot of battle hardened software like resolvd and in doing so it will introduce instability and security issues.

I guess the point is end users will shy away from the short term mess and move to another system like FreeBSD.

While I don't know if SystemD is a vastly superior design and implementation to any of the already existing things and I am not sure how much of a stability/security/complexity concern it is, I think that the 'end' of Linux will not be due to SystemD - it has a lot of momentum going for it. It will take something bigger to derail Linux at this point.

baldfat7• 12 years ago

TL;DR systemd is winning the developer mind share. Things being built like Gnome which require SystemD thus most things will need systemd.

agrover2• 12 years ago

Gnome needs a dbus api that currently only systemd supports, which is not quite the same thing.

andolanra• 12 years ago

It's important to note that the blog post doesn't mean, "Linux is done for," but rather, "Linux is going to be a different kind of system than it has been historically." Some people think this is a good thing, and some think it is not. EDIT: I want to emphasize that I am trying to write this comment neutrally, not endorsing either side, so while my biases may sneak in, my intention is to be fair to both sides. Both sides have arguments in their favour, and I'd hope people consider both before coming to a conclusion.

All Unix-based systems start by running a single process known as 'init' which is responsible for setting up the system, starting all other programs, and managing various services as they run. The change under contention is the widespread inclusion of a relatively new piece of software called systemd, which replaces the historically popular sysvinit, Ubuntu's alternative known as upstart, and various other competing systems. All of these are different approaches to building an init system.

In contrast to some of the other systems, systemd is written less in terms of traditional Unix-style tools (like pipes and plain text files for storage), choosing instead to build on newer and more elaborate communication interfaces and store in specially designed binary formats. Systemd consists of a large family of interrelated pieces of software, many of which are nominally optional but generally expect to be used together. One point of contention is whether systemd is "too large", as proponents argue that developing these pieces together will increase their quality, while detractors argue that this makes it too difficult to substitute components if necessary and that these pieces should not be part of the same conceptual package.

Additionally, the architectural choices of a large, widely-used package integral to the functioning of a running system will influence the design and assumptions of other pieces of software and even Linux itself. Already, the Linux kernel has incorporated the newer communication systems used by systemd into the operating system itself (the mentioned KDBus), which many believe to be a sign that the inclusion of systemd will change the way that Linux operates and the way programs expect to interact with the kernel and with each other.

An important factor here, whether good or bad, is that these choices make Linux into a very different system than it has been historically, and very different than other Unix variants. Proponents argue that this is a step forward, as the facilities offered by Linux historically might not be appropriate tools for the current uses of Linux. Detractors disagree. Either way, the conclusion is that systemd is a change in the way the operating system is structured and used which has been a major point of contention in some communities.

why-el• 12 years ago

I know we don't have access to the OSX code in its entirety, but could one argue that they actually did change how init works in FreeBSD? And perhaps that is one of the reasons the OS is stable? It's not far to assume that the Linux folks are going for more control over init for specifically this reason.

gnufied3• 12 years ago

If you read over Wikipedia page of systemd, you will find out - it just does not replace `init` system, but whole lot more. Systemd is a collection of 64 binaries which manage login daemon, networking (DHCP) etc.

In other words to some it seems like lot of not-so-tested software replacing software that was well tested. from an outsider perspective two things seem problematic:

1. A lot of Linux software worked on principle of, don't break userland. Systemd appears to break userland here and there.

2. From any such large software replacing a well tested infrastructure, bugs are expected. The problem appears to be, in many cases systemd developers push the blame of breakage to other subsystem devs (sometimes it could be Kernel dev, sometimes it could be end user apps written on top of KDE/GNOME). This is the part which makes lot of people angry apparently.

johnny2210• 12 years ago

it was tested enough to ship in redhat enterprise linux 7.

noobermin• 12 years ago

It has been said before, but it deserves mention, there are alternatives to Systemd out there, like OpenRC which has been around for a while, and other, newer projects like Uselessd.

Of course, whether those other options will be as well supported and developed as systemd remains to be seen.

gnufied3• 12 years ago

I am no expert on this but apparently when Gentoo developers tried integrating OpenRC they ran into too many bugs.

http://blogs.gnome.org/ovitters/2013/09/25/gnome-and-loginds...

simoncion8• 12 years ago

...apparently when Gentoo developers tried integrating OpenRC they ran into too many bugs.

You are certainly no expert on this.

I've been using Gentoo since 1.4... back in 2003, maybe 2004. I vaguely remember when Stable Gentoo was switched to OpenRC, which happened in mid 2011. Unstable Gentoo (which I ran -and still run- on my laptop) switched to OpenRC much earlier [but that date I cannot remember].

What the first topic in the blog post that you linked to is really saying is:

"We Gnome developers still claim that recent Gnomes don't require systemd. We stand by this assertion. Recent Gnomes only require init and system management daemons that behave exactly like systemd in pretty much every aspect; they don't actually require you to have systemd installed.

It's a pity that developers don't want to spend time re-working their init and system management software to be systemd clones. If they did, then the world would finally understand why we Gnome developers continue to assert that Gnome doesn't require systemd."

pnathan4• 12 years ago

I am very happy with my Gentoo system and OpenRC - it works extremely well. Modulo jokes about "burn-in test OS". :)

vezzy-fnord2• 12 years ago

There's been at least ten or so sysvinit alternatives that have existed before systemd. Some newer ones have appeared since then, most promising of which is nosh.

That said, all of the sysvinit replacements back then (eINIT, initng, depinit, s6, perp, etc.) never made it and were all ignored during their time.

andyl• 12 years ago

Everything I read about SystemD is negative. Negative on the technology, negative on the people who created it. Nothing positive.

How is it that SystemD is about to dominate the market? Who is driving SystemD adoption, and why?

toolz4• 12 years ago

Why is everything on the news always negative? Is everyone a crime peddling pedophile rapist murderer drug addict? My point is that negative drama sells headlines.

johnny2210• 12 years ago

people like me. I've been waiting for something like systemd to come to linux since i started using it (back in 2000).

Tired of having to write init scripts for every distro, writing forking daemons, and dealing with the too simple syslog.

kbuck• 12 years ago

It hinges upon the fact that people are more likely to be vocal when they don't like something than when they do. Consider online reviews: are you more likely to review a product when it's worked fine as you expected or when it died a month after you got it?

systemd has both good aspects (e.g. faster boot times and removal of the nasty nest of shell scripts) as well as bad ones (incredibly monolithic). There are arguments on both sides of the fence. It's just that you're more likely to hear from people who dislike something than you are to hear from people that like it. (And at this point most of the systemd proponents have given up talking about it because of the incredibly vocal and relentless opposition.)

jqm2• 12 years ago

"removal of the nasty nest of shell scripts"

What one person finds nasty another may find sexy.

simoncion8• 12 years ago

(e.g. faster boot times and removal of the nasty nest of shell scripts)

OpenRC does this quite nicely. Gentoo has been using it as the default init since at least 2011.

baldfat7• 12 years ago

Linux and SystemD work and work good. Because the people who use it like it and we don't have many complaints besides logs are no longer text files we have to use journalctl. The strong critics usually go the route of "It remains to be seen if this is a good or bad thing."

Systemd has been a replacement for the worst part of Linux, init. It was confusing and was just barely workable. It has been needing replacement for 5 or more years. There were SEVERAL solutions brought out and well more developers like SystemD more then the rest.

As a user in Arch Linux, OpenSUSE and Fedora it has been rock solid and I have been able to do things at my knowledge level much more consistent and lower level.

zobzu4• 12 years ago

you dont have to use journald. you can send back from journald to syslog (eg rsyslog) and that works well/its easy to do. You can even disable the journald logging itself and just have rsyslog (or the logger of your choice).

at my place we kept both journald and have rsyslog to redirect logs to the network. ie our /var/log is relatively empty (rsyslog isnt setup to log to disk) so we look at log on the machine via journald if needed (rarely - since we look at the central log aggregator instead)

rbanffy• 12 years ago

Everything I read about SystemD is negative

May be a bias in action. People who like it, or who really don't care much about it (I dislike journalctl, but, apart from that, I'm mostly OK with it) don't waste time writing how not much changed for them and how things continue to work as expected.

mickt• 12 years ago

There does appear to be a lot of negativity around systemd, such as people giving out about binary logging. I'm sure there's got to be a way to enable text logging along with binary logging.

This lad seems really enthusiastic about systemd logging and systemctl:

http://0pointer.de/blog/projects/journalctl.html

There are a lot of interesting features, but personally I'd prefer to have both binary and text logging. Text logging is in cases where a system goes tits up and you may only have access to some basic tools such as grep, vi etc.

Mikeb85• 12 years ago

It's so terrible (literally the devil incarnate, sacrifice your offspring now) that it's already been adopted by:

- Arch Linux

- Fedora/RHEL

- openSUSE/SLES

- Mageia

- NixOS

- CoreOS

- Sabayon

- Debian and Ubuntu in the near future

Everything I read about SystemD is negative. Negative on the technology, negative on the people who created it. Nothing positive.

There's a vocal group that seem to think it's some Red Hat conspiracy to destroy Linux and take over the world. You don't hear the positives because people either don't care about init systems (as long as their distro continues to work) and happy programmers are generally silent on the issue.

The fact that the article says that Solaris is the 'best of breed stack right now' and also suggests CoreOS (apparently not knowing they use Systemd) should speak volumes...

pnathan4• 12 years ago

Well, in fairness, I did have a RH chap tell me to my face several weeks ago that Linux needs to replicate the Windows registry model and stop using the /etc text file approach. So I have a certain sympathy for the idea of Red Hat conspiracy (what is a company anyway but a conspiracy to produce work & products for hire?).

rodgerd5• 12 years ago

The fact that the article says that Solaris is the 'best of breed stack right now'

That well-known paragon of SysV systems 8)

jude-8• 12 years ago

One thing that needs further investigation is the degree to which distros adopted systemd because they'd rather not maintain patches for udev, logind, KDE, GNOME, Wayland, etc. to keep them decoupled.

I have a hunch that for !RedHat, the motivation to switch comes just as much from not having to fight upstream (if not more so) than from systemd's relative merits. After all, there are plenty of different init systems, but only one of them is coupled to the rest of userspace to a significant degree.

Also, I wouldn't write off Solaris zones :) They're pretty powerful--in some ways, more so than LXC is today (syscall translation comes to mind).

lokedhs• 12 years ago

Solaris does not use systemd. It uses systemctl. I'll leave it to others to debate their individual merits, but I personally like the Solaris system a lot.

anonbanker5• 12 years ago

Conspiracy: a group of people working in secret to commit a crime.

In this case, we have red hat, lennart, Kay, and freedesktop.org working together to force a LGPL RPC loophole to circumvent the GPL. Sounds like an appropriate term.

pnathan4• 12 years ago

My impression is that a very small number of people (core maintainers) actually drive adoption of technologies - persuade them & you have altered the cruise ship.

For me, I plan to stay as far away from systemd as I can; it's an abomination of software design.

mrweasel• 12 years ago

I didn't really like systemd initially, but know I'm more in the what-ever-its-fine camp.

Companies like RedHat clearly drives systemd adoption, because it solve problems for their customers, that would be large enterprise customers. Many of those who dislike systemd has simpler requirements, so systemd becomes some new they need to learn, but it doesn't provide them with any tangible benefits. They didn't see or care about many of the problem or features that systemd address.

Of cause you're going to be negative if a new system, one you didn't need is forced upon you. If you're happy with it you're most likely a large company that doesn't blog about systemd, especially if it's something that just works.

At least that's my take on it.

lmm6• 12 years ago

How is it that SystemD is about to dominate the market? Who is driving SystemD adoption, and why?

Redhat, Gnome and Pottering. While no-one can know exactly what they're thinking, it's in Redhat (and other commercial linux vendors') interest for more software to be more tightly coupled to linux and not run on other OSes. And Gnome is mostly developed by people who work for these commercial linux vendors (in a notable contrast to KDE, which has a wider spread of contributors, including more hobbyists and a number who are supported by government grants, particularly in Europe. Which leads to a different set of incentives).

My impression is that Gnome and associated software is, by and large, the wedge which is driving adoption in other distributions like Debian; previous "faster init" solutions were often available as options but never pushed to default because it wasn't necessary; users who wanted a different init could install one, but all their software would work fine either way.

the_why_of_y8• 12 years ago

Of course the good people of KDE will take a stand against the evil corporations and would never ever consider using that vile systemd-logind abomination!

http://blog.martin-graesslin.com/blog/2014/10/libinput-integ...

general_failure8• 12 years ago

Its the same way Dropbox was trivial, iPad is useless and so on. Once the non devs adopt it, you will see the positively side.

imanaccount2475• 12 years ago

Redhat has more influence than people believed. Systemd is entirely driven by redhat, who controls a bunch of other software and said "we're breaking all the software we have so it only works with systemd". Distro maintainers would have immense amount of work on their hands forking everything to try to avoid systemd.

felixgallo• 12 years ago

It's necessary to understand that desktop linux has been total shit always, and this has caused incredible pain over the years to people who, against all reason, continue to try to use it. It turns out the bazaar can't create a polished user-friendly product, period.

Nevertheless, Red Hat and Canonical and others continue to try to foster desktop linux, on the now-obvious misguided theory that (a) the desktop matters and (b) they can take share from windows and osx installations.

As a result, there's hundreds of terrible paid desktop programmers, and thousands of their users, who are dying every day because, e.g., the last several iterations of wireless networking scripts were written by morons, their graphics libraries are comically bad, etc., etc.

Into this charged mix of total incompetence and frustration comes a small group of mediocre coders with hubris, backing, and political nous. And what they are promising to the long-abused desktop users sounds amazing to them, like wizardly magic, and literally, and almost entirely, boils down to this: freedom from having to deal with the shitty wireless networking script system. No joke. That is the fundamental issue at play, and the driver behind "faster boot times", "socket activation", and all of the other marketing points. If that idiot who wrote the wireless provisioning scripts had been competent, this entire mess would never have happened.

So the desktop linux users and desktop linux developers, who again have been living in a tiny cage being pooped on every day by their own regrettable choices, reach for this solution with the religious fervor of a drowning victim. And since desktop linux developers tend to be the C team, they don't care about good architecture, they just want things to work for them and their very specific desktop linux use case, which objectively and axiomatically, again, has not worked for decades and will never work.

So they band together in unison, following the exciting, energetic, charismatic and opinionated lead developer. And obviously, Red Hat is delighted, because that's their employee, and maybe they get more market share. And they pack Debian with developers, because there's a ton of horrible little desktop linux apps that grant them votes, and shout down the opposition. And they set up an IRC channel, and they brigade every forum with the same nonsensical attacks on the very architecture that made it possible for the internet to happen in the first place. Including, obviously, this very thread.

In actuality, this group of users is vanishingly tiny compared with the linux installed base, which is mostly phones and servers, where the real action is, and which don't need this halfassed dbus nonsense or the accreting blob of carelessly rewritten known-good-daemons. The desktop linux people are chasing a dead target with a terrible design and religious fervor substituted for technical ability. It will be intriguing to watch it play out. FreeBSD is about to get a big positive jolt of people that know what 'good' is.

peterwwillis2• 12 years ago

How is it that SystemD is about to dominate the market?

Religiously passionate salesmen, selling an upgrade to your car stereo, who conveniently forget to mention that you also have to replace your whole car.

Who is driving SystemD adoption,

Kay Sievers, Harald Hoyer, Daniel Mack, Tom Gundersen, David Herrmann, and its creator, Lennart Poettering [1].

and why?

One guy wanted to make booting his desktop faster. [2]

Later it was decided that systemd would be "a big opportunity for Linux standardization. Since it standardizes many interfaces of the system that previously have been differing on every distribution, on every implementation, adopting it helps to work against the balkanization of the Linux interfaces. Choosing systemd means redefining more closely what the Linux platform is about." [3] Basically, they want to change "how we put together Linux systems." [4]

[1] https://en.wikipedia.org/wiki/Lennart_Poettering [2] http://0pointer.de/blog/projects/systemd.html [3] http://0pointer.de/blog/projects/why.html [4] http://0pointer.net/blog/revisiting-how-we-put-together-linu...

davexunit4• 12 years ago

I'm quite happy to be a part of the development team for GNU Guix, a distro that is not using systemd. I'm not a systemd hater, but it's definitely not for me and I'm not thrilled with the direction that development is going. It's a shame that sysvinit and friends are so bad that using systemd is the best option we have right now. Maybe GNU dmd will be able to stand up to it someday.

Shish2k• 12 years ago

I'm quite happy to be a part of the development team for GNU Guix, a distro that is not using systemd.

Thank you for doing this :) I love systemd myself, but I still think it's important to have alternatives available; also it makes me very happy to see somebody creating their own choice instead of tearing down other people's choices :D

TheLoneWolfling3• 12 years ago

Thank you for doing this.

That being said, it's a little too esoteric for my tastes (among other things: "if you are looking for a stable production system that respects your freedom as a computer user, a good solution at this point is to consider one of more established GNU/Linux distributions.").

Do you know of any Linux distros that a) don't use systemd, b) are vaguely active / supported, and c) run on UEFI? (My current laptop, unfortunately, has UEFI. It's a royal pain, but oh well.)

davexunit4• 12 years ago

it's a little too esoteric for my tastes

Yes, we're still in alpha. Not ready for prime time yet.

Do you know of any Linux distros that a) don't use systemd, b) are vaguely active / supported, and c) run on UEFI?

Gentoo? I use Debian most of the time, which of course uses systemd now.

josteink2• 12 years ago

Do you know of any Linux distros that a) don't use systemd, b) are vaguely active / supported, and c) run on UEFI?

Not addressing any other of your points, doesn't most laptops/computers which ships with UEFI allow you to set it to boot in "legacy-BIOS" mode?

Even if you're currently UEFI-booting, I would be seriously surprised if UEFI-support was a requirement for every OS your machine can boot.

anonbanker5• 12 years ago

I'm looking forward to switching to Guix as soon as it reaches beta. All this POSIX breakage and LGPL exploitation is making me double down on GNU.

davexunit4• 12 years ago

Glad you are interested! If you are brave, you can try out the distro and report the issues you run into. It would be very helpful to us. BTW, I'm typing this from my standalone Guix machine. Eating my own dog food.

derengel• 12 years ago

Are there any reliable VPS providers that offer FreeBSD?

stock_toaster• 12 years ago

There are quite a few. Some decent ones I have heard good things about:

* rootbsd.net * arp networks * vultr

There are quite a few cheap-ish dedicated server vendors out there that can either be ordered with FreeBSD or provide some means to install it yourself.

PurplePanda• 12 years ago

I have had good luck with arpnetworks. So far only noticed one outage in three years of use. Have had good network and CPU performance. They don't do tech support for things inside your operating system, but they are always helpful when I ask them to do things on their end.

snksnk• 12 years ago

I would recommend http://www.ramnode.com/

Reliable and cheap, even more so if you take 5 minutes to google for a discount coupon.

dijit2• 12 years ago

tilaa.com for EU servers, I like them a lot- IRC support and a general "no bullshit" feel.

imanaccount2475• 12 years ago

I reccomend vultr. I run OpenBSD not FreeBSD, but it is just a standard KVM VM you can run whatever you want on, and unlike most BSD offering VPS providers, they aren't double the price of DO and friends.

zobzu4• 12 years ago

we've already upgraded to RHEL7 here. ppl didnt know systemd too well but it works (im the only guy who knew systemd. heck ppl didnt even know of journald).

Now im not saying that i like systemd in general - for the reasons the author explains I don't really like it. Systemd could have been way better if it wasnt for the political shit and attempts to take control over distros, kernel, etc.

That said it has some features I do like. Likewise for kdbus. This is like Chrome vs Firefox. People would like to support Firefox better. But Chrome and Google apps do enough of what they want right now to go with the solution they don't really approve of - I'll say it again: it works.

sandGorgon3• 12 years ago

single parent hierarchy for namespaces

Predictably, all the blame is laid at systemd's feet.

The current churn is happening, because all of Linux's core developers (kernel and user space) are wanting that change...to push the envelope.

For example, the current change in CGroup's namespaces are because kernel is mandating that the current cgroup access mechanism be deprecated. They want a single writer to Cgroups. Systemd is in the unfortunate position of complying with that request. Guess what? Soon enough, so will Upstart.

Again with kdbus, the person who made the push is not "evil" Lennart, but Kay Sievers - a long time maintainer of udev.

Systemd is nice. Don't be afraid.

Http://www.lambdacurry.com/systemd-nice-dont-afraid/

jude-8• 12 years ago

I'm pretty sure it's Greg KH who's running kdbus these days [0]. At least, he's the one who submitted the patch to lkml [1].

[0] https://github.com/gregkh/kdbus [1] https://lkml.org/lkml/2014/10/29/854

nisa4• 12 years ago

You mean this Kay Sievers? http://www.theregister.co.uk/2014/04/05/torvalds_sievers_dus... - It's basically the buddy of Lennart at Red Hat - at least that is my impression from far away..

systemd may be nice but it's coming from Redhat and cgroups are changed due to because systemd folks wanted it that way as far as I followed that debate...

sandGorgon3• 12 years ago

Yes, we are veering dangerously close to Godwin's Law : not only is Lennart and systemd evil, but everyone who agreed or contributed to it.

johnny2210• 12 years ago

look at how many people actually contribute to systemd. It's far more than just redhat folks

pikzen• 12 years ago

No, cgroups are being changed because the kernel maintainers want to deprecate some stuff and systemd is in the best position to provide the new services.

ploxiln• 12 years ago

If there were a couple of competing alternatives to manage your cgroups, sysadmins might not be so peeved. We're used to installing an alternative syslog daemon or cron daemon rather easily. But it's systemd or the highway, in the 3 or 4 most popular distros.

About udev, Linus has had multiple serious complaints about udev maintainership since GregKH passed it to Kay. Don't you recall the async firmware loading issue...

rodgerd5• 12 years ago

Here's a though: try remembering you don't speak for "sysadmins", you speak for yourself.

pnathan4• 12 years ago

No. Rebuilding the entire userspace set of services to be a systemd cluster is not nice. It's essentially redoing the traditional Linux approach, which has worked relatively well for years. There are a number of things that could be split out and made more modular - c.f., uselessd for a more in-depth analysis.

To be clear, I'm not claiming that SysV init is The Best Way. Shell scripts are not the Happiest Place. But I am claiming that systemd is a crummy and overbearing replacement.

pgeorgi• 12 years ago

It's essentially redoing the traditional Linux approach

It seems like that's part of their mission statement, given comments like this: "Some day, we will have turned the old crap into a real operating system. :)" -- Kay Sievers (https://plus.google.com/+TomGundersen/posts/eztZWbwmxM8)

Andys2• 12 years ago

Freebsd has jails which is a much more baked technology and implementation than LXC

Disagree!

1. FreeBSD has nothing like cgroups or namespaces. You can't apply cpu or memory limits to a whole jail, only individual processes in that jail.

2. it is early days for virtual network cards and ethernet bridging and jails: you have to recompile kernel to add VIMAGE.

Sanddancer• 12 years ago

You've been able to apply cpu and memory limits to entire jails since 10.0 was released through the rctl mechanism.

Andys2• 12 years ago

Thanks. I don't think it changes my point - various attempts were made to implement this feature since Freebsd 6 - almost 9 years ago, for some reason these were incomplete or ignored.

FreeBSD 10 was released only this year. For reference, Heroku began in 2007, using resource limits with LXC... a whopping 7 years ago.

AaronFriel• 12 years ago

Perhaps this is a controversial idea, but is this not just someone finally taking the tried and true Open Source "advice" to heart?

That is, every time I've reported something is broken, wonky, doesn't work reliably, et cetera, I've been told, "Submit a patch.", "Write some code.", or worse, "Implement it yourself."

Someone finally got fed up with the haphazard state of affairs in Linux-land. Fed up with the fragmented and sometimes many places you have to look for error logs. Fed up with the many files you have to edit to configure the network correctly (different on every major distribution). Fed up with the half dozen ways to configure X, where X is a common function to every modern operating system.

It seems Lennart has taken the advice and followed through, and distribution maintainers liked it. They liked the idea that someone was taking all this complicated work - this dirty, boring to write and maintain code - and making their lives easier. Why else would nearly every distribution be on board?

Systemd is offering a more compelling solution than anyone else, and if you don't like it, well, you should submit a patch, write some code, or implement it yourself.

badgersandjam20• 12 years ago

I had no opinion on systemd until yesterday. In fact I had a glace or two at the code and it's pretty clean and I liked the rough objectives laid out.

I installed CentOS 7 on a machine last night that we're replacing CentOS 6 on and was poked in the face with timedatectl and dbus problems for an entire hour, some of which were intermittent. Debugging these issues is a horrific pain. I lost 4 hours on it. I've never lost that much time on a system function before. This is not what I expected and there is no way I could possibly introduce that to our production environment.

I think that might why people are slightly sensitive to it.

Yes you're exactly right, but replacing something with something less stable, more complicated and more difficult to debug isn't a rational or good engineering. I'm sure many people will be fed up with systemd much quicker than what was already there.

Not impressed with a community which pushes this as stable, quality software. Voting with my feet: FreeBSD is being trialled instead. WhatsApp throwing a million dollars at it draws a lot of valuable attention and puts it in the business's mindset.

Choice is as much of a valuable aspect of open source too...

paulfurtado• 12 years ago

Out of curiosity, what ended up being the root cause of the timedatectl problem?

gtirloni• 12 years ago

It hardly does anything for FreeBSD in the enterprise world. Companies cannot afford to support Windows Server, Linux AND another flavor of UNIX (FreeBSD). They are already dumping Solaris/AIX/HP-UX as much as they can so environments are a bit more homogeneous (and easier to support). There is no point in onboarding FreeBSD for what they perceive as minor technical differences (that, truth be told, are overshadowed by dozens of layers of bureaucracy and any efficiency benefit is completely lost).

It took a long time and a giant ecosystem to get where Linux is today at big enterprises. OSes are commodities in that space. They are not commodities in many other spaces though (e.g. startups, HPC, science, etc).

olau• 12 years ago

Are you really extrapolating from ONE data point?

In some sense systemd is more stable in that it's fixing some longstanding bugs with sysvinit, but of course it will have some bugs of it own. If you don't want to deal with that, you could skip a release.

nailer• 12 years ago

I installed CentOS 7, with no prior systemd experience, had no problems and found .service files way neater than sysvinit bash scripts and liked having meaningful names in the log/ journal rather than 'local3'.

billybofh• 12 years ago

I've been a longtime linux sysadmin (back into the 90's) and have run into similar issues as you. I've never had as many problems with basic system stuff as I have since testing some of the new systemd-defaulting distro's. For instance, the centos/rhel 7 boxes I've tried have erratic problems - sometimes not setting the hostname, sometimes services don't start, sometimes services I've disabled _do_ start. It's making me really think about shifting to FreeBSD (or, god help me back to Solaris).

seivadmas• 12 years ago

You know this is funny, I remember reading comments EXACTLY like this about 3-4 years ago but with pulseaudio in place of systemd.

Pulseaudio was Lennart's previous project. It broke everything in linux sound for a while, everybody moaned and hated it and said it was the worst thing since the crucifixion of Christ.

Yet, name one problem you had with sound on linux in the past year? There are very few. Pulseaudio now just works(tm) and is a unseed, unheard of part of the plumbing.

If you remember what is was like messing with ALSA and (shudders) OSS before pulseaudio came along you will agree that the current state of affairs is a million miles better. It used to be really difficult to get more than one application to be able to play sound at a time. I remember compiling sound drivers from source just to get them working. Configuring ALSA config files to get surround sound working was practically a black art. Creating manual scripts that unmute the sound card on every boot because the driver didn't initialize it properly.

With pulseaudio, I never have to worry about any of that and configuring surround sound takes me two clicks of the mouse.

Lennart did a fantastic job with pulseaudio, he took on a dirty problem that nobody else dared to touch and went through years of criticism to produce a really high quality solution that solved the linux audio problem so well that you don't hear complaints about it anymore.

In light of that, I trust him to do a good job with systemd. It'll be a couple of years of everyone moaning and bitching and whining about it, then one day it will have become a seamless part of the plumbing, everyone will take it for granted and wonder how they ever managed fighting with shell scripts and fragmented init systems before systemd came along.

It's ironic that Lennart Poettering is probably the most abused developer in the entire OSS ecosystem, yet he is one of the people contributing most to it. For our sake, I'm glad he has such a thick skin. If I was him I'd have quit this game long ago.

zackelan• 12 years ago

I completely agree with you that Lennart scratched an itch, which is the way all good software gets started, and others picked up on it.

Where I think the systemd-naysayers have a valid point is around the tight coupling that has been introduced, and is still being introduced, between systemd and various other components of a fully functional Linux system.

To take your "just submit a patch" example - say N years from now I'm unhappy with some aspect of how systemd works. I can submit a patch, or I can rewrite that whole component from scratch. However, it's entirely possible that the piece I'm unhappy with is so tightly coupled to the rest of systemd that I can't rewrite one component of it without rewriting the rest of systemd, or convincing the systemd maintainers to accept my rewrite and bake it in as the new "official" version of that component.

Where I think the criticism of systemd is valid is that the idea of modularity has taken a backseat, and the APIs between the different components of systemd haven't been very well-thought-out. The informal spec is "whatever systemd does today is correct", which of course destroys any sort of interoperability.

And by way of full disclosure, I'm an Arch user, and run systemd on 4 systems I use everyday - home desktop, home server, work desktop, work laptop. Whatever else I have to say about its design, I use it every day, and actually like the parts of it that I use. eg, the boot time for my desktop is stupidly fast, and if I want to know about some log message, I just run journalctl. I no longer care whether the foo daemon uses syslog, or writes to its own /var/log/foo.log that I should set up rotation for, or handles its own rotation as /var/log/foo/2014-11-20.log, and so on.

And just to play devil's advocate with my own position - there's a certain point where tight coupling makes sense. Linux kernel modules, for example, are tightly coupled to the Linux kernel, and don't work unmodified when compiled against a *BSD or Solaris kernel.

hwh• 12 years ago

Well, the very distro-specific bunch of scripts in /etc/init.d (or is it /etc/rc.d/init.d/? or /etc/rc.d? or a symlink to ...?) were some kind of tight coupling, too.

Plus: This tight coupling did not exactly replace existing communication features. It created new ones. These are made use of.

Yes, systemd is bringing lots of new functionality. Under the hood - that is why sysadmins love or loathe it and users mostly don't care. That "tight integration" argument is mostly one that comes from people (please do not take offence, you're weighting it carefully indeed!) who bemoan that other userspace system infrastructure is left behind feature-wise. And those who love to argue about and against design decisions.

rythie• 12 years ago

You can't realistically submit a patch to change the direction that Systemd's is going in. For example, they won't accept a patch which removes 95% of the code, so a more modular system can be built.

Submitting a patch, implies you agree with the general direction but need a bug fixed or a feature added.

XorNot4• 12 years ago

You can fork it. You can fork all of Debian. But there's numerous reasons that's going nowhere.

Most being: this is not the issue the loudest voices say it is.

pdkl955• 12 years ago

Not only would submitting a patch be agreeing to their goals (Lennart gets to push the Overton window a bit further), suggesting that we should simply submit patches presupposes that the systemd cabal would ever accept them. Unless it is perfectly in agreement with their goals - including the complete software - they probably won't accept it. They don't even accept already written and tested patches for trivial things like #ifdef-ing a couple minor fixes so the project can build on a different libc.

Lennart Poettering[1]:

    Humm, I know this will disappoint you, but we are not particularly
    interested in merging patches supporting other libcs, if those are not
    compatible with glibc. We don't want the compatibility kludges in
    systemd, and if a libc which claims to be compatible with glibc actually
    is not, then this should really be fixed in the libc, not worked around
    in systemd.

If they aren't interested in trivial compatibility patches, they certainly aren't going to accept any patch that dares to disrupt their tight integration or questionable design choices.

As for forking the whole thing, remember that logind was briefly liberated so it could be build as a standalone package, Lennart went and did a big rewrite so the next version was much more integrated with systemd. When he controls the internal APIs and can change them whenever he wants, a clone will have to be a total replacement right from the start, or it ends up perpetually having to catch up to the changes that will be introduced just to cause breakage.

[1] http://lists.freedesktop.org/archives/systemd-devel/2011-Jul...

pjc50• 12 years ago

Or stop using Linux entirely and move to FreeBSD.

icantthinkofone4• 12 years ago

Linux Torvalds said he would never have created Linux is FreeBSD was available at the time. And the only reason it wasn't available is because it was embroiled in a lawsuit over embedded ATT UNIX code back in the 1990s(?).

erglkjahlkh• 12 years ago

Unfortunately FreeBSD doesn't have very good MAC system. Capsicum and friends implement only ~10% of what SELinux offers. Also the virtualization is far more basic, and lacks tooling (it's like receiving a huge bag of Legos and instructions on how to build space shuttle).

jbb555• 12 years ago

Systemd is offering a more compelling solution than anyone else, and if you don't like it, well, you should submit a patch, write some code, or implement it yourself.

Or go back to using windows for general use and software development which is in fact what I've done. It's amazingly sad after many years of being a strong linux supported but this has killed linux for me. I see no point in continuing to use it.

badgersandjam20• 12 years ago

That's really not a terrible idea actually. I switch back and forth a lot between the Windows and the Unixy ecosystems and by far the best end user experience, tailored[1] developer experience and (I'm going to get slated for this), mobile experience is on Windows by far.

[1] tailored as in platform specific. For cross platform stuff, it's not so good.

lmm6• 12 years ago

Lennart didn't just submit the code and put it out there. He lobbied other projects to hard-depend on it, and lobbied distros to adopt it. Systemd didn't succeed where less poisonous equivalents failed because it was technically superior (it isn't), it succeeded because of shady back-room politics.

mdellabitta• 12 years ago

How dare he ask other people to use the software he wrote.

ngrilly2• 12 years ago

Let's say you're right. What kind of "political" leverage Lennard had to convince distros to use systemd?

peterwwillis2• 12 years ago

Why else would nearly every distribution be on board?

This informal fallacy is based on the idea that everything in the world can be distilled to a single answer. The real answer is more complicated. For example:

Red Hat is on board because they pay its creator's salary. So they rely on an individual bias.

Debian is on board because they rely upon 'collective wisdom' and committees to make decisions. So they rely on the bias of group thinking.

Ubuntu is on board because Debian is on board. So they rely on the bias of the other.

Other distributions are using it because 'every other distribution is using it', or they're small enough that it doesn't cause conflicts for its use base, or because it's a GNOME dependency, or because it's just new technology.

To make someone think something is a good idea, show them someone else thinks it's a good idea. This is a fact of all human beings' thought processes. Decisions are not based on merit, or logic, or even a quorum; it is based on fallacies created by heuristics. There exists a heuristic in which the more an idea is adopted, the more other people think it is a good idea.

We imagine our thoughts are logical, and that other people also think logically, and that their decisions must be made for a good reason. But in fact, the great majority of all decisions we make are based on guesses; this is how our brains are able to carry out complex calculations and come to decisions in split-seconds.

For example, you might look at systemd and say, "it fixes so many problems! it provides so many features! it standardizes Linux! CLEARLY this is superior. we must adopt it."

For people who care about the purity of the highest technical ideals, this makes sense. For people who care about being able to use their computer, these things don't matter, and in the systemd implementation, actually makes things worse for them. The changes systemd purports to make are not bad things. It's really just the way in which they did it that is bad.

It's like wanting to upgrade your bicycle to four wheels, but requiring the rider now operate it lying flat down and using mirrors to navigate. The four wheels was a great idea. Using mirrors to navigate? Maybe not so great.

Of course, its creators will turn this inconvenience into a feature, saying "you get to lay down! it's therefore more efficient and easier to use!", completely ignoring how other people want to ride a 4-wheeler.

blutgens• 12 years ago

haphazard? fragmented? many places for error logs? half-a-dozen ways to configure X?

Have you ever run a linux box? How about dozens or hundreds of them in a production environment? I'm guessing no on both counts based on the nonsense you're spewing forth in your post.

If you don't know what you're talking about, it's best to just keep quiet.

101914• 12 years ago

"Someone finally got fed up with the haphazard state of affairs in Linux-land"

But wouldn't the path of least resistance be to switch to a project that does not have this "haphazard state of affairs"?

When I originally tried Linux I got fed up within _days_. It is the _relative_ lack of default "configuration" (that is decided by someone else) that makes me stay with FreeBSD and NetBSD. Of course, lack of default configuration is the antithesis of popular Linux distributions. Whenever I have to use one, I spend more time learning how to turn things off than I ever did learning how to turn things on.

The answer to the original question is, I think, "no", switching is probably not the path of least resistance for many Linux users. Because when the Linux user makes that switch, they immediately find that someone has not done everything for them.

And from what I have seen, observing the questions of Linux users who first try FreeBSD or NetBSD, they generally do not like that. It means they have to do some configuration of their own. And even if they are comfortable doing configuration, it means they have to learn things that are different from the "Linux way"; and they inevitably encounter shortcomings that are due to lack of developer resources (read: time).

In doing things for yourself you learn about how things work. The rc.d system that all BSD projects use is coherent and relatively easy to understand. For whatever that is worth.

This debate over systemd seems to cut to the core of the value of learning about how things work. The reader can draw their own conclusions.

Linux is only a kernel and it should still be possible and thus optional to run that kernel with a basic init (or init alternative, e.g., one based on daemontools) and with userland utilties that do not need systemd.

The question I have is how difficult the popular Linux distribution folks are going to make that for their users to do.

And if they do make it difficult, it raises the question, "Why?"

frozenport2• 12 years ago

I used Gentoo for 10 years, and I can't figure out how to get a Systemd configuration working with XFCE. Indeed, most of the existing documentation has greatly slowed me down.

not_with_retard• 12 years ago

That's pretty sad, I was able to get it up-and-running on my first Gentoo installation.

I guess time doesn't equate to knowledge, eh?

frozenport2• 12 years ago

Haha, well its running but not quite right, if it saves me any face it worked out of the box in KDE!

general_failure8• 12 years ago

Fwiw, my friend runs Ubuntu for 10 years and he cannot understand gentoo for the life of him. He thinks it is absurd people want to keep recompiling stuff.

nyir• 12 years ago

People don't want to recompile stuff, they want to configure the software to their liking. Basically it's very enjoyable if you're a developer.

And FWIW with Gentoo/Portage you can create and install binary packages.

SwellJoe6• 12 years ago

The ignorance in this post runs deep.

For example:

Speaking of zones and Solaris, if that’s an option for you it’s probably the best of breed stack right now.

Does the author have no idea what's going on with Solaris? Hint: Nothing. Nothing is going on with Solaris, because Oracle doesn't care about Solaris. They closed the source, and now push out the occasional minor update from on high for their enterprise customers. Anyone who is suggesting Solaris as an alternative to Linux at this point in history is simply not credible.

jodrellblank• 12 years ago

Curious logic you have. "It doesn't have a constant stream of fixes. That proves it's no good. Things that need fixing are better than things that don't need fixing".

SwellJoe6• 12 years ago

Software maintenance matters. A lot. Particularly for networked systems.

Also, it hasn't been Open Source in several years. That makes Solaris utterly irrelevant in my world. This blog post is suggesting Linux users move to Solaris. It's just a bizarre recommendation, made as an off-the-cuff assertion with no reasons given.

Also, Oracle. What hacker chooses to be beholden to Oracle?

bcantrill3• 12 years ago

He's actually suggesting SmartOS[1][2] and OmniOS[3], which are both illumos distros that are very much alive, having forked from OpenSolaris over four years ago.[4]

[1] https://smartos.org/

[2] https://news.ycombinator.com/item?id=8571961

[3] http://omnios.omniti.com/

[4] https://www.youtube.com/watch?v=-zRN7XLCRhc

SwellJoe6• 12 years ago

That's probably a wrong choice, too, for most Linux users (though there are some reasons a reasonable, and technically competent, person might choose an Illumos system, if you're doing it because a random crank on the Internet tells you to, you probably don't know enough to understand those reasons and the quite large tradeoffs you'd be making).

Regardless, it's one example of many where the author exhibits a very poor grasp of...well, everything he talks about. Dunning-Kruger effect is funny that way.

imanaccount2475• 12 years ago

The ignorance in this post runs deep.

The irony there is quite impressive.

Does the author have no idea what's going on with Solaris?

Do you?

http://wiki.illumos.org/display/illumos/illumos+Home

https://smartos.org/

http://openindiana.org/

http://omnios.omniti.com/

Even the commercial Solaris from Oracle is actively updated and a release was done this year. Anyone who is suggesting Solaris is not an alternative to Linux at this point in history is simply not credible.

SwellJoe6• 12 years ago

Yeah, actually, I do. I maintain a software project with a thousand or so Solaris installations.

matthewmacleod5• 12 years ago

Really bored of this nonsense at this point. Take this for example:

If you absolutely MUST run Linux, my recommendation is to minimize the interaction with the base distro as much as possible. CoreOS (when it’s finally baked and production ready) can bring you an LXC based ecosystem

Systemd is absolutely key to how CoreOS works. It's the basis for the distributed init system it provides — a major selling point.

Taking any of this blog's advice would be harmful. I'd suggest a better approach would be to accept that the majority of distributions have settled on systemd and that generally this decision has not been made by idiots. So it would be worth either understanding what their pain points are and how they can be solved with an alternative to Systemd, or to help solve the issues that are apparently in Systemd yourself.

badgersandjam20• 12 years ago

Many large decisions have been made by idiots, badly. Blindly accepting consensus as a rational choice without a counterpoint had caused many years of suffering for the human race from wars to bad science to bad medicine.

Whilst I agree that the blog is probably hocum, there's nothing wrong with critical thinking.

matthewmacleod5• 12 years ago

I completely agree – but critical thinking involves something along the lines of "here are the things that I think are wrong, this is why I think they're wrong, this is how I think they could be solved."

The answer "let's throw everything out" isn't that useful; likewise, dismissing the considered opinion of lots of people who have been doing this sort of thing for a while needs to be done with some rationality. An empty, bandwagon-jumping appeal like this adds little value and just helps spread more misinformation.

devcpp• 12 years ago

Linux developers are generally very smart people in my experience. It's consensus of many experienced and smart people that makes it significant.

vidarh8• 12 years ago

I agree with that specific advice (minimize the interaction with the base distro). I'm on a quest to isolate every major component of my user experience in containers (including things like browser etc.).

But not because I have anything against Systemd. I love Systemd so far.

It's hilarious that he's proposing CoreOS as an alternative, given that it's one of the most radical rethinks of a Linux distro out there.

mbainter3• 12 years ago

For those with reading comprehension skills this isn't funny but rather a natural conclusion. The problem for you is that you're making assumptions about the author that you haven't verified.

The problem here isn't change, or re-thinking linux. The problem is re-inventing the wheel, and doing it poorly.

CoreOS uses systemd, but it's not a distribution in the classic sense -- rather it's a platform for containers. The narrow use-case for systemd here removes some (most? all?) of the concerns.

dgdsgdsg• 12 years ago

for me it seems that most people complaining/crying out loud come from a system admin perspective.

For example the author of this blog article does not seem to have contribute a lot of code to any project and is more "just" an admin and not a dev.

for me it looks like they don't like new stuff love the linux world how they knew it and don't want things to get easier or even change.

icantthinkofone4• 12 years ago

Complaining and crying from a system admin perspective is pretty significant.

hawleyal• 12 years ago

Agree. However, admins and devs often have competing interests. Admins want stability and ease of maintenance. Devs want features, agility, and speed of development. Admins often disregard or downplay technical debt in light of "has been working in production for X years". Devs often don't care enough about backwards compatibility.

edwintorok2• 12 years ago

I don't think they would complain if things became easier. All I hear about is problems with systemd, and having to solve problems (and learn how systemd and all its components interact so you can debug it) is not easier.

general_failure8• 12 years ago

Yeahz they all spend like 20 mins and come out crying without spending at least a week understanding it.

Which software stack do you know that you instantly get in 20 mins. I for one cannot understand the sysvinit or upstart in 20 mins (ie) to develop and debug. How do you debug sysvinit anyway? Echo's I place in the code do not appear on the screen. Can i claim 'Oh shit this is so broken...'?

linuxydave2• 12 years ago

for me it seems that most people complaining/crying out loud come from a system admin perspective.

Sysadmins complaining about it is pretty important, as at the end of the day they are the ones who are maintaining the systems.

Personally, I have a strong dislike for Systemd. I don't believe it's the right replacement for sysvinit however it's not going away so we'll have to try and work around its warts. Also, currently it's buggy as hell so I will use Centos 6/Ubuntu 14.04 for a few more years to see if the many problems are sorted by then.

general_failure8• 12 years ago

That's exactly as intended. People who want to help out systemd will use latestbsoftware, report bugs and fix it because you know thats how free software works.

People who don't want to help out use the most recent stable and have nothing to worry about.

lazyjones2• 12 years ago

I have mixed feelings about systemd: like many others, I've spent a lot of time with traditional Unix systems (25 years) and like the simplicity and stability commonly associated with them and familiar tools/conventions. On the other hand, OS X showed what a modern system built on Unix should feel and look like, while Linux never got anywhere near (not only on the GUI layer). If it takes unreasonable people to make progress (like in the famous G.B. Shaw quote), then let's sit back and see what happens. Perhaps Linux will shed more cruft and become simpler to develop for rather than harder.

Of course, I'm in the favourable position of not having to maintain/administer a bunch of Linux boxes for a living. I can fully understand the frustration of people who built and shipped custom solutions on top of SysV init.

The end of Linux? No, it's the end of Linux as a traditional Unix with lots of arbitrary optional features on top perhaps. We'll get used to it, or switch to something better.

e402• 12 years ago

On the other hand, OS X showed what a modern system built on Unix should feel and look like, while Linux never got anywhere near (not only on the GUI layer).

This feels like a non-sequitur. The nice GUI of Mac OS X has nothing to do with launchd (which is the systemd-like portion of Mac OS X).

scott_karana• 12 years ago

Are you kidding me? Launchd helps coordinate a massive selection of on-demand GUI-related services and their prerequisites. (254 services on my Mavericks install as of writing)

SysV init is ill-suited to that sort of complex event-oriented management, which (for example) is exactly why Canonical developed Upstart in the first place.

chrsw• 12 years ago

Progress is good. But is systemd progress?

lazyjones2• 12 years ago

is systemd progress?

My knee-jerk reaction would be to say "no", but many people who are more involved with this part of Linux seem to think so, I don't feel qualified to challenge that.

I am confident that more radical changes are good in general, as long as the bazaar model allows the best solutions to survive and the worst to be undone if needed.

pdkl955• 12 years ago

Wow. The UNIX-haters are out in force. Apparently proven, real-world tested code is no longer a relevant feature to a lot of people. Well, "Those who do not understand Unix are condemned to reinvent it, poorly."

But enough about the technical distractions. Why is systemd so important? It's certainly not technical quality or design (even if some of the ideas are useful, the implementation is junk). There is a far more important reason, that some of you have noticed parts of, at least tangentially.

From this very thread: "systemd makes a lot of sense for embedded systems". Yes, it does, andor, and it's all because of this: "kdbus: efficient IPC". The thing is, kdbus/dbus isn't really that great for a lot of things - you have to bounce through the kernel at a minimum, and there is and encode/decode steps that add some overhead. It might be useful for some types of IPC, but it is replacing what should be a fast and simple library call in many places.

Now, here's where a lot of you are going to start calling me crazy or "obviously wrong", both without actually addressing the key claim, which was is probably better explained by stevel over in the Gentoo forums[1]. I encourage reading that post.

The goal with all of this is not technology related at all: the systemd takeover is an attempt to separate Linux and many userspace tools from the GPL, so that software can be used under the LGPL terms instead.

What is the big difference between GPL and LGPL? Linkage. Linking to a GPL library requires you to follow certain requirements if you link against it, while the LGPL specifically allows taht usage. (k)dbus provides the workaround, by replacing what would be a normal function call into a library with a "IPC". It's slower, but so what, computers are way faster than needed. In the end, while you can still choose to release your code as GPL, if you have to use an IPC mechanism to do anything useful the license requirements that will actually apply ends up being being more like the LGPL.

Well, if I wanted to release under the LGPL, I would. What I'm not going to do is undermine my choice of license just because a bunch of embedded developers (and others) want to use what were traditionally GPL projects without having to be bound by the coopyleft requirements. If this was proprietary software, you would call that kind of behavior "stealing".

(seriously, the linked comment below does a much better job of explaining this)

[1] http://forums.gentoo.org/viewtopic-p-7645524.html#7645524

icantthinkofone4• 12 years ago

Where do you get UNIX-haters? I read systemd haters which begets Linux haters but UNIX? No.

Besides, Linux is no longer a UNIX-like system. Linux is now only Linux unto itself with UNIX-similarities.

pdkl955• 12 years ago

Just look at this very thread - there are numerous posts trashing the "old ways" of UNIX.

It has been a common thread throughout the systemd mess. Lennart himself is very strongly outspoken against any use of shell scripting (not just in init).

lwh• 12 years ago

It seems like most systemd problems boil down to: shit this will take more than half an hour to understand, I'm a lazy admin and don't feel like learning something new.

icantthinkofone4• 12 years ago

I've been saying some of these things since systemd came out and typically get downvoted into oblivion so here this article makes HN's front page.

His implication about switching to FreeBSD is a good one and I notice a decent influx of Linux admins switching to FreeBSD showing up on every forum I visit.

dschiptsov2• 12 years ago

Going against the "divide et impera" principle, "do just one thing, and do it well", "everything is a text file" (configurations and logs) and "pipes" which are very natural idiom, in a UNIX-like system which has been built upon these principles, is just bad engineering, shallow understanding and, perhaps, too high ambitions of knowing better how to fix what isn't broken.

The question why do we have this project and why there is such a buzz is more interesting.

But the answers doesn't belong to realms of system design or engineering, they are in more obscure notions of "business strategies" and "sales techniques".

It is a text-book example how to sell - "show them that they have a problem they didn't realize before, and give them the one and only fix". Fine. Except that the problem doesn't exist outside discussions of "why this product is good - it solves this and that".

Another thing is that it is an excellent strategy to "create a niche for oneself". Look, we are the ones who give you a solution (to a problem which doesn't exist). We are "world leader" - look, we have lots of experts, web-presence, lines of code.

There are so many examples in "parallel worlds", beginning with eco products or "fair-trade" coffee, you name it.

But you see, lots of people don't need an eco or environmental-friendly daemon which does everything "the right way". No, thank you.

Another thing, back to system design. If Windows has "registry" and there are millions copies of Windows everywhere, or Mac has that settings daemon (from which GNOME Gsettings has been copied) it doesn't mean that these design decisions are superior to plain old text configs. You could compile them and have an "intermediate storage" to gain a 5% more efficiency for application startup times, but this is, again, not a fundamental problem. Being able to use regular expressions on configuration and logs without any restriction are much more fundamental. And memorizig all the details of all these xxxxctl insted of saying something like

   $ grep -R ^sysctl /etc

or whatever I need, is just an additional unnecessary burden.

Most of the problems systemd is trying to "solve" does not exist. At least not BSD, AIX, Solaris, you name it. And, of course, there never been any problem with syslog, init or cron. They were *good-enough".

e12e3• 12 years ago

I don't support a switch to systemd at this time (I use Debian, and I'd be happy to live with pretty much any other init system for at least one more release - the obvious choice would be to just stick to systemV init, and see what systemd looks like in a year or so. Down side is that if the distro were to move (for supporting all those people that want Gnome for some reason ;-) -- then we loose out on a year or two of testing systemd).

Anyway:

Most of the problems systemd is trying to "solve" does not exist. At least not BSD, AIX, Solaris, you name it.

Don't forget that (open) Solaris revamped the init system, not just the disk/filesystem-system. I think there are arguments to be made for integrated systems like the zones/zfs/SMF.

I don't think systemd is a reasonable way to go about it, and I certainly don't think it is a good fit for "Linux in general". My impression is: Systemd is too big -- will fail.

Wasn't it Gnome that at one time tried to mimic the windows registry for settings (because binary data on disk: really fast, lol) -- only to go back to ini-style files? (I might be misremembering that one).

Different designs are fine (see: eg plan9) -- but moving away from fundamental design principles (everything is a (text) file) effectively means abandoning the old system, making a new system.

If systemd was a bit more upfront about "making a new operating system sharing some code with Linux kernel and traditional userland" - rather than trying to sell systemd as "more of the same, just better" -- maybe they'd meet with more positive reception. That a new system is unstable is fine -- just don't expect people to use your new crap in production.

RubyPinch2• 12 years ago

Except that the problem doesn't exist outside discussions of

except https://en.wikipedia.org/wiki/Init#Replacements_for_init

this problem was noticed and was trying to be fixed for a long time, by a lot of people

dschiptsov2• 12 years ago

Traditionally, one of the major drawbacks of init is that it starts tasks serially, waiting for each to finish loading before moving on to the next. When startup processes end up I/O blocked, this can result in long delays during boot.

For a server it doesn't matter, leave alone "broken". long delays during boot is grossly exaggerated, because delays are mostly due to network errors/timeouts or file-system checks, but you can do almost nothing "in parallel" without clean FS or configured network interface.

Well, for some Ubuntu for mobile devices they could have adapt some "optimized" sysvinit-scripts (not /sbin/init) replacement, but, please, don't tell us that everyone needs this. Leave servers and home workstations alone.

the_why_of_y8• 12 years ago

If the problems with sysvinit are all imaginary then why has every other UNIX replaced it?

Mac OS X has launchd for 10 years now.

Solaris 10 replaced sysvinit with SMF.

Don't have any experience with AIX but the documentation says it has a "System Resource Controller" with a "srcmstr" daemon to manage services; though it looks like it runs on top of something sysvinit-like:

http://www-01.ibm.com/support/knowledgecenter/ssw_aix_61/com...

BSDs have generally never used sysvinit, although they do have shell-based service management.

So actually none of the particular examples you cite use purely sysvinit to manage services.

And it's also quite obvious that there are use-cases where reliable and fast service management are very important on servers.

One is containers, where you may run 10000 of them on a single host, and you have to reboot the host some time too, so you don't want to delay the start up of all those containers by loads of slow and racy shell scripts. Or even better, you want to have socket activated containers that don't actually start until they are needed.

https://www.getpantheon.com/blog/pantheon-running-over-50000...

Another is hyperscale servers: There are now really small ARM and x86 based servers and you can put 4-500 of those in a single rack. That means a lot more individual servers to admin, and failure modes that are relatively rare in current server rooms will become an order of magnitude less rare, so more robust OS level service management is helpful.

julie1• 12 years ago

Problem 0: Yes sysinit has problems but systemd brings even more problems (journalctl, PID 1, dependancy Hell ++, undeterministic bootup process)

Systemd is a broken silver bullet for handling the decrease in quality of packaging.

There are problems with sysinit: the first one is the missing link between devs and syadmins in companies: people known as packagers.

For a sysinit to work well shell scripts, permissions, where resources are located, the dependency management has to be done with art and expertise. It is a human job with human which are:

1) a very skilled rare resource, 2) not identified has needed by companies, 3) company's software QA is shitty (it has barely the level it works for me (tm) of quality

Talking about debian, which is considered the distribution with the most talented packagers debian has 2 flows in this domain:

* too rigid: projects goes for logical units of packaging that are consistent and organization of assets in packages that can ease maintaining. Debian has its guideline that makes them «fix» poorly packaged softwares like lateX, python, ruby: cutting language distro in at least runtime, dev, extra. Debian packagers are often debian experts, not upstream software package experts and they first break some stuffs (latex is so poorly packaged on debian it can be considered broken), AND it adds more works

* too much features typical linux distro compared to BSD are pacakging fucking more packages in their core resulting in more work; less attentions to the details and conflicts of functionality/overlaps. This result in more resource drained from the packagers. Like we have 4 shells considered OK for writing shell related stuff, when they have only one: «sh».

The problem of linux vs BSD is symbolized by the systemd vs sysinit: linux is an OS of devops that are super devops, poor coders and sysadmins, BSD is an OS of sysadmins and devs that are good sysadmins and devs, but no devops.

And we still lack in 2014 of maintainer, sysadmins and coders of quality.

Linux/Gnome ... FSF projects are not sustainable in these conditions. They think of free software has an infinite resource of benevolence. And they exponentially overblow the works required for maintaining, deploying ... thus they are mathematically doomed to die under their own weight.

I see BSD as a calvinist boring protestant community turned towards humility doing what is right and linuces as catholic exhuberant rockstars over spending the good will of developers without thinking of the future.

Being lutherian, I still dream of THE right OS that would less terse than BSD.

return0• 12 years ago

As a rule of thumb, it's good to distrust products when you know that the politics behind have been loud enough to be heard widely. But the sharp razor which will decide if systemd is better is simplicity. Those who have switched, is systemd simpler than sysvInit?

jerven5• 12 years ago

Yes, declarative unit files are simpler than Turing complete shell scripts.

sethrin• 12 years ago

Simplicity is not necessarily a good measure of suitability. See this: http://mjg59.dreamwidth.org/2414.html

There are some aspects of systemd that are simpler than sysvinit. The mess of double-forking and PID files goes away, and process management is much more reliable. Setting up containers for services is much simpler. Unit files are generally much simpler than a given init script.

I think more useful adages ("boned wisdom for weak teeth" - AB) would be, "Everything should be as simple as possible, but no simpler," and "Simple things should be simple, complex things should be possible." There is an enormous amount of complexity inherent in the problem of initializing and monitoring a modern OS. If your solution to this is actually simple, it is probably wrong or incomplete. How you manage that complexity and what your abstraction layers are are the vital questions.

morganvachon6• 12 years ago

I think the article got one thing wrong: That sooner or later every GNU/Linux distro will switch to systemd. I can think of three that most likely will not: Slackware, Crux, and Gentoo. Granted, the more that systemd binds itself to formerly modular userspace utilities and apps, the more pressure on those distros to make the switch, or else either fork every systemd userspace app at the risk of being left behind. But I have a feeling there will be a schism in Debian, resulting in a fork or (less likely) dropping systemd altogether, before too much longer. That might just be the catalyst to move away from it in other major distros.

Then again, I could be completely wrong, and systemd will end up in every single surviving distro by default, turning GNU/Linux into systemd/Linux. That's when we'll see a true exodus of those opposed to it to other OSes.

wtallis• 12 years ago

I'm also interested to see how things play out with OpenWRT. High-end home routers are starting to get a lot more powerful with the transition to ARM SoCs and NAND flash, but they won't be able to run any software that depends on the systemd ecosystem while OpenWRT is still supporting the massive install base of hardware that systemd can't fit on.

morganvachon6• 12 years ago

Ditto for small ARM devices like the Raspberry Pi. When the RPi version of Arch Linux went to systemd, the system actually got slower in my experience. On such an already limited device, it was a quantifiable difference; before systemd, Arch was the fastest GNU/Linux distro for the Pi by leaps and bounds. After systemd, even Raspbian feels faster than Arch, and now Slackware is the reigning king of speed for it (again, among Linux distros).

Unfortunately Slackware is still a bugaboo when it comes to installing it on the Pi, but once you've got it on there you can image your SD card and have a ready-to-roll distro that's lean and mean.

mcguire• 12 years ago

Good point. What is the systemd approach to embedded systems in general?

renovaatio2• 12 years ago

Two distros that don't use systemd:

http://www.voidlinux.eu/

http://crux.nu/

Void linux is a rolling release, has binary packages, uses a very good package manager (even better than pacman) and the community is small but very friendly.

Crux is very minimal, and you have to compile your packages.

I use Void Linux, and am very happy with it.

papaf• 12 years ago

Thanks for that information.

I am using Arch at the moment but I spent the last 45 minutes trying to fix a network problem which involved fighting with journalctl to see what was actually happening.

Void linux looks like a great alternative.

renovaatio2• 12 years ago

I assure you void is a great aternative to Arch. I was with arch until systemd, and it feels like home, only smaller, with Arch.

phkahler• 12 years ago

The description of systemd - monolithic, taking on function of other things - sounds like the X server. Look what's happening now with Wayland, and look at how hard it's been to get there (hint Wayland is only possible as a result of all the driver architecture changes over the last 5-8 years). It's hard to undo this kind of thing once it's entrenched, but it can be done.

e76204• 12 years ago

Look how many platforms X supports, how many paradigm shifts has survived. systemd already has a less flexible architecture by design that an ancient system, so in the future it'll be difficult to untangle.

SEJeff2• 12 years ago

He rants about systemd, then talks about liking CoreOS, which uses systemd for everything. He is contradicting himself, but I'm genuinely not sure of why.

angersock3• 12 years ago

Read it again:

Author is saying that CoreOS is a good solution because you use it with heavily isolated containers. Thus, any use of systemd is unable to screw up hosted applications. For that use case, it makes sense, whereas in more general use, systemd wants to get all up in everything.

timf2• 12 years ago

He responds to that here as well: http://blog.lusis.org/blog/2014/11/21/a-few-things/

saurabhnanda• 12 years ago

New to this entire brouhaha, but I'm concerned because my entire production stack runs on Linux. At some point devops is not going to be able to do things the old way.

Is there a link where one can read about the rationale of why systemd was needed? Perhaps by the person who started off the initial project?

scourge2• 12 years ago

is this a joke? - "my entire production..." you're a devops engineer? - "link where one can read about the rational.." with good grammar? - "why systemd was needed... " ok you lost me there - "the person who started off the initial project?" and you've been hidden under a rock for the past 10 years??

http://0pointer.net/blog/ - is his blog

are you seriously trying to tell me that someone is going to have to explain what a "container" is by comparing with chroots? :D how have you been running your production stack all these years if you didn't have an internet connection under your rock :)

sciurus• 12 years ago

https://news.ycombinator.com/item?id=8641839 is a much needed response to this discussion.

exabrial• 12 years ago

The real reason people are pissed about systemd is it's intentionally designed to take over the entire system. Conveniently is also maintained by RedHat. So yes, it is a hostile takeover of Linux.

timf2• 12 years ago

The author responds to a number of the comments here: http://blog.lusis.org/blog/2014/11/21/a-few-things/

willhamina• 12 years ago

Does systemd increase the security of the OS or reduce it? For example, does it allow currently-separated processes to dip into each others' memory space? I would appreciate a birds-eye view from someone familiar with the security ramifications of systemd.

Would not the incorporation of many loosely-coupled but individually secure mechanisms into a single monolithic mechanism be useful to an entity whose purpose was to monitor communications, view/modify systems unbeknownst to sysadmins and users, etc.? Yes, I'm talking about the NSA et al. I reference the following which also brings up Red Hat's control of Linux:

"Julian Assange: Debian Is Owned By The NSA"

http://igurublog.wordpress.com/2014/04/08/julian-assange-deb...

orionblastar2• 12 years ago

I think that SystemD is an attempt to put Linux on par with Mac OSX and Windows 8.X, and attempts to do that haven't worked in the past.

Like Canonical changing Ubuntu to use Unity and Mir, so that it was easier to use on tablets and modern PCs. It just doesn't seem to work and it drove me to Lubuntu and use LDXE with an XP like Start Menu UI.

Ubuntu doesn't use SystemD yet, but I got a feeling it will.

I am downloading Fedora 20 because it has SystemD in it, so I get some idea what it is like.

But I think every Linux company wants to become another Apple for some reason. They saw how *BSD Unix went into making Mac OSX, and they want to try and copy that with their Linux distro and right now this SystemD seems like a path to that.

It is like a change from free and open source to commercial Linux. We all saw how Lindows/Linspire tried that and failed.

I might go to ReactOS, HaikuOS, AROS when one of them gets finished to the right level that it is ready for prime time. All they need to do is port Apache2, PHP, MySQL and other stuff to those operating systems to be used in VPS hosts so they can be used as alternative to Linux with SystemD if it ever breaks things.

pekk2• 12 years ago

Which Ubuntu release did you find to be using Mir as a default?

orionblastar2• 12 years ago

I think it was on a tablet, not on the desktop by default yet.

It is kind of confusing.

sztanpet• 12 years ago

You know the blog is just a clickbait when it capitalizes the name of the project wrong.

(on the wiki, the big Spelling header could not be more explicit)

Systemd redux: The end of Linux

Related Stories