Debian Talks About Systemd Once Again

An anonymous reader writes: A couple of months ago the technical committee for Debian decided in favor of systemd. This is now a subject for discussion once again, and Ian Jackson says he wants a general resolution, so every developer within the Debian project can decide. After a short time, the required amount of supporters was reached, and the discussion can start once again.

Comments

Some Sense Restored?

[Anonymous Coward]

Hopefully they'll come to their senses and reject the disease that Pottering has cursed the Linux ecosystem with.

Re:Some Sense Restored?

[FudRucker]

i have to agree, i was running Jessie for a while and systemD is just awful, i rather have the old init system that Debian has used for years, it works and i figure if it isn't broken don't fix it, i have since removed Jessie and switched to Slackware and have made up my mind to stick with Slack and keep my eye out for using only non systemD distros

Re:Some Sense Restored?

[CRCulver]

Debian is a lost cause. From Gnome 3 to Systemd they've lost their independence.

Debian's offering of Gnome 3 and Systemd are not comparable. Gnome 3 is only the default desktop for people who just want to click through the installer. But you can completely avoid Gnome 3, and indeed many people do because they do e.g. headless server installations or choose another GUI. Systemd, however, is spreading through so much of the system that it may not be possible to avoiding installing it. Even if one hangs on to sysvinit as one's init system, programs like Gimp on Debian now end up pulling in systemd libs.

Re:Some Sense Restored?

[mlts]

I personally would like to see it (and its evil compatriot, firewalld) as options.

In RHEL 7 and downstreams, you can choose between LVM2, standard partitioning, or btrfs as ways to carve up your disks. It would be nice to have systemd as an option, so for laptops where parallel starting of daemons makes a nice speed increase, it is useful. For a server where one doesn't want many changes to the underlying OS unless it is something to be tested, it can be an option. If one is using containers, maybe systemd might be useful to have.

There are changes to Linux like SELinux and AppArmor which are must haves. These add significantly to the security of the OS. systemd does add security... but not really that much. One can specify that a program run with ulimits and possibly in a container, but a wrapper can do the same thing, and one thing that I get concerned about is one program having so many moving parts that touch everything on the system, even perhaps the TTY functions.

Re:Some Sense Restored?

[Anonymous Coward]

If the moving "toward the 21st century" means suffering an all-encompassing virus that diverges entirely from the established design philosophy of the context it resides in then get me a ghetto blaster, some jolt cola and my stone washed jeans because I'm living in the 80's forever.

Re:Some Sense Restored?

[csnydermvpsoft]

The problem with supporting multiple init systems is that each package that provides a daemon needs to support all of them. A traditional init script is just a shell script, while upstart and systemd have their own formats. You could write software to convert an upstart or systemd script to a shell script, but there would likely be cases where it wouldn't be easy to translate automatically.

With filesystems, applications don't need to know anything about what's mounted how and where—you could mount /var on a btrfs partition on LVM2, /home over NFS, /tmp on an ext2 ramdisk, /usr on a read-only CD-ROM, /etc on a floppy... and everything would just work (albeit slowly because of some of my hypothetical choices).

Re:Some Sense Restored?

[goulo]

That Gimp (and more and more other programs) require systemd (which is supposedly merely "an init system") is the really evil thing. Poettering & Red hat are explicitly trying to force all Linuxes to have systemd.

I don't trust Poettering & Co's track record, nor competence, nor intentions (e.g. seeing him use sleazy manipulative rhetoric in a conference video where he accused a systemd opponent of not caring about handicapped people), and I sure don't want their unnecessary huge mass of dubious code on my machine. (Though I'm sure the NSA will be happy for the increased opportunities of exploits in such a huge messy mass of code.) And even if this "init system" were somehow really necessary for Gnome, I don't use Gnome.

They have lied, e.g. claiming that systemd is just an init system, or that it is not a big monolithic chunk of code, yet it becomes more and more monolithic. E.g. I just watched a week or two ago as several libsystemd packages in debian became merged into a single package while one user was trying to create a stub for one of them to satisfy some needless systemd dependencies by some applications.

I am becoming increasingly convinced that Ignorant Guru is right, and Linux is being manipulated for corporate interests, not users' interests. http://igurublog.wordpress.com...

Re:Some Sense Restored?

[sjames]

The funny part is that systemd has nothing to do with making a good desktop system with things papered over. Once the whole cgroups kernel interface will be stabilized, I would expect any number of improvements on the SysV init to take place.

Start with the parallel init already available in Wheezy and add a simple daemon manager called in the init scripts to stick a system daemon in a cgroup and manage it and you have every advantage systemd offered and none of the drawbacks.

If desired, that manager could support the "I'm ready" callback through a passed FD (a pipe endpoint). For non-Linux systems, the wrapper can support the callback and skip cgroups.

My big concern over systemd hasn't been that SysV would go away, but that a mediocre at best replacement would wedge itself in through crazy dependencies and prevent the better solution from even starting.

Re:Some Sense Restored?

[gweihir]

Actually, there is no problem with systemd as long as you can avoid it easily.The fanbois shall have their toy, I have no issue with that. But forcing it on everybody, even those that actually have a clue about good and reliable system design, is just wrong.

Re:Some Sense Restored?

[gweihir]

Well, obviously "moving to the 21th century" means being ignorant about things that work. New is not equal to "better" in any way. Maybe they could put a Farcebook client into systemd as well, that would show the quality level of the design of this thing clearly.

Re:Some Sense Restored?

[thegarbz]

That people think GIMP and GNOME require systemd is outright absurd. They both depend on a single feature exposed by the kernel which have nothing to do with the init system. It just so happens that the most prevalent API available for this is currently the logind component of systemd.

Rather than bitch and moan about them tying the two together, why not start / sponsor the start / donate to an alternative API that's not part of systemd which GNOME / GIMP can depend on for the functionality they need.

As for Poettering's track record. His software is released early in it's infancy, that and that alone (in my minority opinion) is his big problem. All of his previous projects have resulted from a very real need to clean up some of Linux's most stupid (again in my opinion) design features. People like talking about the disaster of pulse-audio, but those same people have never had the fun of attempting to plugin a USB headset to take a call or transfer audio to another device currently playing, or never had to try and get bluetooth audio work. For all it's complaints pulse-audio is now mature and (in my opinion) works rather well.

systemd is not just an init system. The only people who claim that are those that haven't understood what Poettering is making. It's a complete system management platform. I have no opinion on if it will be good or bad to go this route, but it does look like it will solve some very real gripes that I and others have with the current Linux setups, which includes the arcane task of digging through log files. (Ok I have an opinion that binary logs aren't the way to go, but the old system was just screwed).

Re:Some Sense Restored?

[CRCulver]

You know what's funny, you blast systemd as insecure yet it is every init based system with Bash installed (most of them, though certainly not all)

Debian (you know, the topic of this discussion) and its many derived distros like Ubuntu had for many years used dash as their /bin/sh. Init wasn't passing anything to bash. Indeed, the switch was made because dash offered faster boot times for a script-based init system than bash.

Hope!

[Anrego]

A very well written proposal that outlines many of the concerns I (as a non-Debian user) and I suspect most have about systemd. It’s worming it’s way into everything for the sake of better integration, which it may deliver on, but this goes against much of the traditional Linux spirit of small self-contained bits that can be swapped out at will.

In my mind, this comes down to whether we want a better functioning OS or an OS that adheres to the mindset that I think attracted many of us to Linux in the first place. Personally I want a hackers OS that I can play with and tweak as I feel like, but I accept that many people basically want open source windows or even just zero cost windows (i.e. free as in my wallet).

I hope Debian rolls back on their decision. I doubt this will happen, but at least we’ll get some more discussion in a somewhat visible forum. I may not agree with a lot of the Debian mentality, but they are very good at thinking about and discussing things, so I think this will be good overall.

And before someone says "just use gentoo", I do, and have for almost a decade (I started using it fairly soon after it came out). The problem is that systemd, being basically a virus at this point, is causing exactly the kind of problems mentioned in the proposal. I've had to use the blacklist for the first time in a while because *McBane voice* the use flags, they do nothing!

Re:Hope!

[FudRucker]

maybe systemD will cause a Fork in the Debian distro,

i like Slackware too, maybe Slackian_Debware

Gentoo users will make a Debtoo

Re:Hope!

[rjmx]

> In my mind, this comes down to whether we want a better functioning OS or an OS that adheres to the mindset that I think attracted many of us to Linux in the first place.

I'm not even convinced that it makes for a better-functioning OS. I've been a Debian user for 12 years, mostly running 'testing' distributions. When systemd first turned up, I let it run for a couple of weeks, but switched back to sysV after half of my startup daemons didn't. Tried it again a month or two later, but when it had trouble stopping Samba (and, worse, claiming that it would wait *five* *minutes* before killing the processes, I decided enough was enough, and now I'm in the process of switching all five of my Debian boxes to Gentoo. Now, granted, the testing distribution is for just that purpose -- testing -- but if I'm dealing with the kind of idiot that would claim that systemd results in faster startups, but thinks that a five-minute wait to shut down a process is acceptable, then I want no part of it.

Debian should listen to what its users want, rather than just its developers. We're not all technically ignorant.

Re:Hope!

[MightyMartian]

Binary logs are anti-*nix. Rebut that.

Re:Hope!

[Sarten-X]

this goes against much of the traditional Linux spirit of small self-contained bits that can be swapped out at will.

In my mind, this comes down to whether we want a better functioning OS or an OS that adheres to the mindset that I think attracted many of us to Linux in the first place.

Personally, that principle of having many swappable self-contained bits is one of the worst qualities on UNIX.

I've been using GNU/Linux for over a decade. I know my way around most distros, and I can usually figure out what I need to do to accomplish any task... usually. The biggest problem I face now is that distros have so many small components doing their small tasks that figuring out exactly which component is responsible for a given task is no small feat.

I understand and appreciate the programming simplicity that a small component brings, but from a user's (or admin's) perspective, the operating environment is now more cluttered. As distros pick and choose their preferred swappable components, the view gets worse. Sure, I know exactly what the "finger" command does, but it's not obvious that "pinky" is an alternative, because having a lightweight finger command is apparently an important thing. Some distros will even create symlinks or scripts to provide alternative common names for their chosen packages, but there's seldom a guarantee that the input or output will be the same. This is why the first step of many build processes is to examine the environment and figure out exactly what is available on the system, often using methods that uncomfortably remind me of browser-detecting JavaScript.

I'm not saying that systemd is the solution we need, or even that it is a solution. I've just dealt with far too many poorly-named packages to have excessive reverence for this archaic principle.

We should also keep in mind that Linux itself, as a monolithic kernel, defies the concept. By design, the kernel's one job is to interface with every piece of hardware on the machine. Is it really so far out of line to define systemd's one job as interfacing with every service provider in the OS?

Re:Hope!

[Anrego]

I've used gentoo for a long damn time, so my ability to objectively gauge it's difficulty is probably long gone.

That said, I for one think gentoo has gotten far easier to install and especially maintain. The default profiles are no longer the joke they once were, and most packages are using more generic high-level use flags so you have one --with-feature-x instead of the old --with-compat-mode-z --with-doublefork --with-some-other-unrelated-but-required-flag type stuff you had years ago, which translates into much simpler USE flags. You can actually leave make.conf relatively untouched and still end up with a decently functional system, especially if you want a desktop and go for one of the desktop profiles.

Portage is also a lot smarter these days, being able to resolve many issues that it previously would have died on. When it does run into problems, the descriptions these days are much nicer than before!

I'm being completely honest when I say that systemd has been the first major gentoo headache I've had in a while. Everything was just dandy then suddenly I'm having to switch packages around (udev being the big one), and having to blacklist udev and systemd because so much random shit pulls them in (and a -systemd use flag isn't enough), and then uninstalling a bunch of random packages (like some power management widget that got pulled in by god knows what for some reason).

I know you've probably written off gentoo at this point, here's a completely random bit of usage advice:

- Set use flags as you need them, even if this means re-installing the same thing multiple times. This avoids big important packages being pulled in as mere dependencies (though you can add them to the world list afterwards) and more importantly lets you set up and configure everything one at a time and makes it more likely that you'll notice error messages.
- Don't be afraid of package.keywords, especially for very specific use flags.
- Avoid gnome if possible. I don't know wtf it is with gnome, but it seems to be the poster child for weird and hard to diagnose issues as well as crazy dependency trees.
- Pay attention to what virtual packages are doing. Usually they are in your best interest, but not always.
- Don't bother using ebuilds for web apps

Re:Hope!

[Anonymous Coward]

Simple. She wants the D and is willing to use the System to get it.

Re:Hope!

[bill_mcgonigle]

And...that is where binary logging should stay, until it can be eliminated entirely.

I don't even know why I'd care if systemd uses a binary representation internally, as long as it can give me human-readable logs with only text tools.

Only showing binary logs with systemd tools is a misfeature of the type "exposing the implementation". Userland requires a UI, and it's bad UI, and frankly bad Unix.

Now then, I hear you can somehow configure systemd to echo a copy of its logs to rsyslog. But, and maybe I'm just a fool with poor GoogleFu, but I tried for a couple hours to get this working and only found company for misery on the mailing lists.

If any systemd fans can point us to a quick-n-easy HOWTO on getting text [r]syslog working under systemd, then by all means shut a few of us up. Tell us how there's plenty of documentation too, we'll all hang our heads and wander away.

Re:Hope!

[Anonymous Coward]

And yet all the complaints ignore the fact that systemd is a diverse modular set of tools, and not a monolithic blop, that can work perfectly well with other tools (like the usual system log).

Consisting of a boatload of different programs which are all required to work with each other and may not be switched out is not modular; it is, in fact, nearly the very definition of monolithic as applied to software.

And that not much code actually runs in PID 1.

Define not that much? How much is too much if the design and implementation are deeply flawed?

And that the TC found in its favor.

Railroaded by Red Hat, as anyone who actually followed the process knows.

And that before systemd there was a default init system that everyone used as well.

Because everyone was able to agree that that one worked in the way it should.

I'm not qualified to judge the technical merits of the case...

And yet you do.

...but the people that are overwhelmingly came down in favor of systemd.

Comments attempting to support systemd always seem to trot out this little chestnut, despite the fact that there is no evidence whatsoever to support it. It's almost like there's a playbook somewhere telling you all what to say.

The criticisms on philosophical grounds were rebuted and the rebutals stand unopposed.

Rebuttal: You keep using that word. I do not think it means what you think it means.

Instead the same questionable objections are raised again and again....

Which in most sane groups would indicate that the objections are perhaps not so questionable. Nice editorializing, though.

Re:Hope!

[Peter H.S.]

Binary logs are anti-*nix. Rebut that.

That is of course wrong. If you have a POSIX compliant system, you have binary logs in /var/log. On Linux they are usually called "utmp" and "wtmp" and they keep track of logins and logoffs. You use the "last" tool to read these binary logfiles. utmpx is actually a formal part of Unix.

Remove It

[Anonymous Coward]

Debian is by far the most stable of the Linux distros. systemd does not lend itself to this stability. Nothing wrong with the old init system. We all know it and its quirlks. I fell in love with UNIX because of editable text config files. Every aspect of the system needs to be editable by an admin. Linux is losing morally to OpenBSD because OpenBSD does not allow binary blobs in the system. Ever. Debian should be the same. No binary blobs of any kind. If it's not text, it doesn't belong.

Re:Remove It

[rubycodez]

Not FUD, if something wrong you'll never get to the part where you forward to syslog. Logs should be simple text files, that can be read even without the OS. ASCII text is viewable on just about anything

Re:Remove It

[Anonymous Coward]

Systemd represents the ongoing Redhatification of Gnu/Linux. Wonder why all those stupid complex looking projects come from Red Hat ? Although they cannot claim ownership of Linux, they can make it so that most Linux components are guided by Red Hat. Which is the next best thing as far as they're concerned. As for Linux users ? Who the fuck cares about them.

Re:Remove It

[Anonymous Coward]

I think in your rush to tell people they've missed the point completely, you've missed it yourself.

If you're reliant on trusting the logs of a system that you think might have been compromised you're already shafted. If an intruder can edit your plain text logs then they can edit everything else on the system as well, including binary ones; hacking is generally more sophisticated than vim /var/log/daemon.log dd dd dd :wq. There's nothing inherently unhackable about binary logs and if your box is rooted your only real option is to burn the hard drives to the ground and reinstall.

"Just add some hashes! Then the hash won't match and the log will reveal itself as being tampered with" comes the chorus. Nope. If you've already got the ability to rewrite the logs then recalculating the hashes so they match the massaged data is trivial.

The only remotely secure way to record logs is to write them off to a seperate log server (firewalled off, no remote access, managed by a completely different team, barbed wire, attack dogs, Slim Whitman, yadda yadda) the second they're generated - although by all means keep a local copy if you feel like it.

Windows has the whole binary logging thing, and goes to great pains to make them as hard to access as possible (which of course makes viewing them a pain as well). But anyone with admin access can clear the logs or delete whatever entries they like with winzapper or a few lines of code. Windows in secure environments uses... wait for it... a remote log server.

Long and the short of it is binary logs don't get you any extra security and, especially in the nix world where there are many and varied tools for examining plaintext logs, constitute a colossal loss of readability.

Disclaimer: I work in computer forensics. Most hacks are done by people who haven't even thought of covering their tracks and you'll have nice local log entries that tally up with those on the remote server that scream out "Look, here's me hacking teh gibson!". Advanced hacks are almost impossible to spot without a) a good IDS b) examining the discs offline.

Re:Remove It

[Lord Apathy]

Oh horse shit! Binary logs are the biggest crock of shit to blow out someones ass since Windows ME. Pure text log files can be parsed with the same system tools that come with all linux distributions. I'm talking grep, awk, and sed. An you really adventurous, perl.

All the excuses that I've seen for keeping this foul system are bullshit!

Re:Remove It

[Lord Apathy]

Exactly. Text log files allow you to pull them off the damaged system and examine them on another system. Many times the system that you have to examine them on won't be a linux system. The other day I had to examine a log file that my co worker sent to me on my phone.

If it had been a binary log file, I would have had to get off the toilet, get dress, and find a compatible system to decode the log files on. Maybe even spin up a VM to do it in.

I was able to tell her the boot params to boot up the box with out leaving my seat. Fuck binary log files.

Re:Remove It

[Lord Apathy]

yes, so export the binary file to text files.. not difficult

A completely unnecessary step.

An one that might be impossible if the server has promptly stuck its thumb up its ass and won't boot. Then add that you are a junior system admin that doesnt' know how to do that. But you know enough to boot the system with a rescue disk so you can mount the drive and then ftp the log files some place where you can mail them to your boss at 2 am. Who might just reach through the network and strangle you because you took down his porn server anyway.

Not everything speaks binary log files, but just about everything speaks ascii. Log files are in ascii on purpose. So you can read them with the minimal tools and not have to fight the system to get to them.

Just because you can do something, doesn't mean you should. That applies to binary log files.

Please Debian

[Anonymous Coward]

I've been a Debian user for 14 years now, please do the right thing and get rid of systemD.

I've been trying systemD on another machine for about a month now, it's not terrible but it's not all it's cracked up to be either.

The part that I don't like (besides it going against the unix philosophy) is how fast it's taking over before the majority of the Linux community even had a chance to have their say. And what really gets me is, if systemd was just an init system, fine. But at the rate they are going there is going to be a systemd everything.

Completely wrong

[Anonymous Coward]

The summary is completely wrong. They are not discussing systemd, just whether packages can depend on a specific init system. I thought there was some kind of moderation here?

Re:Completely wrong

[Anonymous Coward]

A key point is the systemd approach to things seems to directly contradict this goal. It's seems almost by design to be getting hooks into as much as it possibly can to make removal very difficult. It's the lamprey eel of init systems.

In a world where GIMP, a graphics editing tool, has a dependency on a specific init system, it's hard not to discuss whether this was a good idea in the first place when discussing the replacability of that init system.

I'm hoping this is the path these discussions go down. Continuing to support systemd is going to lead to a two tiered Linux where not using systemd excludes you from a tonne of software, and this is about as anti-Linux as you can get.

One of the worst points about systemd

[NotInHere]

is for me that it isn't interoperable. Please correct me when I'm wrong, but AFAIK systemd never did anything to create standards their new functionality is compatible with. Instead they only support linux APIs. I recognize that their needs exceed POSIX, but their current approach "lets make everything a hard dependency" is -to be polite- hacky. It doesn't have to be an official ISO standard, a simple document that ensures exchangeability of components inside systemd, and perhaps even makes systemd cross-platform.

Re:One of the worst points about systemd

[Anonymous Coward]

This is one of my major gripes as well. I think if we're going to start rewriting/updating software to the spec of a better init system, it needs to be a universal specification that many (perhaps systemd-like, perhaps not) alternative init systems of the future can adhere to, include those on other *BSDs and *nixes and operating systems we haven't even dreamed of yet.

I really dread a future in which, due to current Linux dominance and all distros going systemd, all of the major software packages start depending on systemd's APIs and behaviors, and then the software packages become very hard to port sideways or forwards to other platforms. Don't get me wrong, I love Linux. However, what I love more is the idea and culture behind Linux and all *nix/*BSD. I want there to be alternatives, and I want there to be future upstarts that disrupt Linux and give us something even better.

The reason the culture of all of this was so disrupt-able in the first place, leading to all the greatness we have today, is because we had *standards*, and new implementations could adhere to those standards and all the other software could quickly be ported over to them.

Aside from technological gripes about how systemd operates and/or is implemented, the key failing of systemd is failing to specify a standard for authors of everyday runtime software and daemons, so that those other authors can conform to that standard, and then the *BSDs and whoever else in the world can implement that newer, better standard independently of systemd.

Systemd is like an anti-standard. They seem to have never even *thought* about it from a standards perspective. They think only in a functional perspective, and the only function that seems to matter to them is "Today's current iteration of Desktop Linux systems". Arguably even within that limited realm systemd has standards issues. They make incompatible changes from release to release and hardly even mention them in their changelogs, much less provide backwards compatibility or a path for sane future feature changes.

OpenRC

[Anonymous Coward]

Seriously, why not OpenRC?

It solves all the deficiencies with classic init, but at the same time it doesn't have the interoperability problems and un-Unix-like feel of systemd.

Maybe it's just time for RedHat ReactOS

[morgauxo]

Maybe Potering and his other buddies at RedHat are great, the best thing since sliced bread.... but... they are working on the wrong OS. These guys don't belong in Linux. They belong working on ReactOS! Imagine ReactOS with all of RedHat's resources behind it. It could quickly be a better Windows than Microsoft's! Meanwhile those of us who like Linux as Unix and aren't in the market for "Free Windows" can go on enjoying a better Unix than Unix. We could all be happy!

Ian Jackson

[inhuman_4]

For those who don't know, Ian Jackson was the most vocal anti-systemd proponent on the committee. Considering that last time systemd was up for vote he tried: strategic voting, usurping the committee chairman, and finally throwing a temper-tantrum and refusing to talk to anyone for a few days. When it was all over he promised to try and reverse the committees decision with a General Resolution.

And now having failed to win on technical merits, he is back at it again trying to kill systemd via 'loose coupling'. Something that the committee declined to rule on.

Re:Init is broken

[Anonymous Coward]

It can't do event driven launches and yes it does impact desktop users. Example, your on your corporate network on a laptop and you close the lid and take a plane to somewhere else and the laptop wakes up. How can Init handle something like this and know to configure it to a new network?

This is why Sun, Apple, and Ubuntu developed their own event driven systems. System D is not good. But event driven systems can respond to events like a hack attack, excess load, and other things for servers.

Init was made for stationary mini computers with only 20 text based commands and apps. It's not designed for the hacks we use to get it to work today on modern systems

So why would/should that be a function of an init system, and not a function of, say, an event driven network daemon that gets started by init on bootup - and receives event notifications on you opening/closing the laptop lid (suspend/resume)? Why should that need an entirely re-written init system when it could easily be done in just a (rewritten) network daemon using the existing init system?

Sounds to me more like entirely missing the point of what the init system is supposed to do... which is *not* reconfiguring NICs, etc (those are functions of the scripts/daemons it starts).