LWN: Comments on "Debian looks at OpenRC" https://lwn.net/Articles/512719/ This is a special feed containing comments posted to the individual LWN article titled "Debian looks at OpenRC". en-us Thu, 16 Oct 2025 02:06:07 +0000 Thu, 16 Oct 2025 02:06:07 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net Debian looks at OpenRC https://lwn.net/Articles/514703/ https://lwn.net/Articles/514703/ mpr22 <blockquote> The (binary) interdependencies are reaching an all-or-nothing level, and for some folks it's by design. I grudgingly half-accepted /bin/ls dependincy on libselinux (why?</blockquote> <p>Presumably to get <tt>getfilecon()</tt> so that it can support the <tt>--context</tt> command-line option. (The alternative would be to have <tt>/bin/ls</tt> invoke the dynamic linker, which doesn't seem a terribly pleasant alternative.)</p> Mon, 03 Sep 2012 11:35:00 +0000 Debian looks at OpenRC https://lwn.net/Articles/514668/ https://lwn.net/Articles/514668/ jackb <blockquote> I'm on the verge of giving up on binary distros. The (binary) interdependencies are reaching an all-or-nothing level, and for some folks it's by design. I grudgingly half-accepted /bin/ls dependincy on libselinux (why? -- and no, "it's just a teeny-weeny lib", wielded down-thread doesn't convince me. It doesn't seem to scale).</blockquote>You should take a look at Gentoo. USE="-selinux" solves that problem. Sun, 02 Sep 2012 15:39:18 +0000 Debian looks at OpenRC https://lwn.net/Articles/514665/ https://lwn.net/Articles/514665/ oldtomas <div class="FormattedComment"> Thanks, rleigh, for putting things so clearly. I couldn't have.<br> <p> To illustrate: I'd managed to set up a Debian system without dbus (libdbus had to go in -- even Emacs depends on it these days). One upgrade and *whoosh* this thing wants to go in[1] (i just could avert it this time by ditching recommends).<br> <p> I'm on the verge of giving up on binary distros. The (binary) interdependencies are reaching an all-or-nothing level, and for some folks it's by design. I grudgingly half-accepted /bin/ls dependincy on libselinux (why? -- and no, "it's just a teeny-weeny lib", wielded down-thread doesn't convince me. It doesn't seem to scale).<br> <p> Time for a "reduced" Debian derivative, perhaps: "Curmudgeon's Cut"?<br> <p> --<br> [1] among other goodies with the suffix "-kit"<br> </div> Sun, 02 Sep 2012 11:20:45 +0000 Debian looks at OpenRC https://lwn.net/Articles/514557/ https://lwn.net/Articles/514557/ HelloWorld <div class="FormattedComment"> <font class="QuotedText">&gt; e.g. I see no reason for them</font><br> So what? They won't interfere with your work if you don't need them, the overhead is negligible, so please, just name *one* sensible reason for disabling cgroups. Btw, file permissions are also totally useless on my wireless router, so why doesn't anybody complain because you can't disable those? Because they don't harm anything either. <br> <p> OTOH, there is a reason for enabling cgroups they're needed to reliably terminate a service, including all its children. Perhaps you don't need that, but many others (like Cyberax) like that ability.<br> <p> <font class="QuotedText">&gt; ...but <a href="http://secunia.com/advisories/search/?search=dbus">http://secunia.com/advisories/search/?search=dbus</a></font><br> Yeah. So how many of those redundant are reports because dozens of different distros ship the same fix for the same problem? How many of them affect d-bus-related stuff that isn't required for systemd (dbus-daemon, dbus-qt, dbus-glib etc.)? How many are made irrelevant by the fact that you need root permissions to even connect to systemd's sockets? Can you even name *one* real problem, that was caused by systemd's use of dbus, security or otherwise?<br> <p> <font class="QuotedText">&gt; Do you know the difference between PID 1 and a bunch of other processes, Mr. Downright Downplayer?</font><br> rleigh's original argument was that systemd shouldn't make use of cgroups because then it's hard to remove/modify them if they turn out to be a bad idea. In that context, whether the PID is 1 or something else doesn't make any difference whatsoever.<br> <p> <font class="QuotedText">&gt; What about accepting patches?</font><br> What for? As I said, other operating systems already have their own init systems and it's unlikely they'll switch to systemd (the BSDs aren't exactly famous for loving the LGPL). OTOH, making systemd portable will likely compromise its maintainability. It's all pain and no gain.<br> <p> <font class="QuotedText">&gt; What is irritating in the crowd touting systemd is that viral "who would need this or that?".</font><br> systemd can do everything sysvinit does and more. Nobody stops you from launching a shell script from a systemd unit if that is really what it takes, but systemd units cover the vast majority of cases just fine and they're easier to write and not as ridiculously inefficient as shell scripts.<br> <p> <font class="QuotedText">&gt; we thoroughly dislike an idea of running systemd at servers.</font><br> And yet you completely failed to name any sensible reason for that conviction.<br> </div> Fri, 31 Aug 2012 20:16:00 +0000 Debian looks at OpenRC https://lwn.net/Articles/514556/ https://lwn.net/Articles/514556/ moltonel <div class="FormattedComment"> <font class="QuotedText">&gt; On the other hand, it would mean somebody with a stake in BSD would</font><br> <font class="QuotedText">&gt; have to do it, and that is unlikely to happen because of NIH.</font><br> <p> While that (may be) true of the FreeBSD/OpenBSD/etc crowd, it's obviously not the case of the Gentoo/BSD and Debian/BSD one.<br> </div> Fri, 31 Aug 2012 18:30:27 +0000 Debian looks at OpenRC https://lwn.net/Articles/514390/ https://lwn.net/Articles/514390/ Cyberax <div class="FormattedComment"> <font class="QuotedText">&gt; e.g. I see no reason for them (and the system in question isn't the one where stray processes are welcome in the first place).</font><br> So you prefer to blindly trust all daemon developers that they correctly process all the shutdown requests and won't hang your box during restart?<br> <p> For example, BIND had a bug where it hangs indefinitely on exit if DNSSEC is used. Doubly nice on a remote server, especially if init kills sshd first. And yes, that caused me a real-life 4am trip to our datacenter to press the "reset" switch.<br> <p> <font class="QuotedText">&gt;&gt; There's also nothing wrong with systemd's use of D-Bus.</font><br> <font class="QuotedText">&gt;...but <a href="http://secunia.com/advisories/search/?search=dbus">http://secunia.com/advisories/search/?search=dbus</a></font><br> Systemd uses dbus for communication between trusted processes as a simple form of RPC. It can run totally fine without system-level bus.<br> <p> <font class="QuotedText">&gt;&gt; thus making systemd portable would be a waste of time.</font><br> <font class="QuotedText">&gt;What about accepting patches?</font><br> Has anyone actually bothered to create a compat patch for systemd?<br> <p> <font class="QuotedText">&gt;Heck, I've suggested this article in the distro's devel mailing list just in case OpenRC might get considered and found useful -- we thoroughly dislike an idea of running systemd at servers.</font><br> Why? Because of various phobias?<br> </div> Thu, 30 Aug 2012 18:24:47 +0000 Debian looks at OpenRC https://lwn.net/Articles/514379/ https://lwn.net/Articles/514379/ gvy <div class="FormattedComment"> <font class="QuotedText">&gt; I can't help noticing though that the successful desktop operating</font><br> <font class="QuotedText">&gt; systems do apply a well-integrated approach.</font><br> In marketing (and sometimes in plain deception), that is.<br> <p> <font class="QuotedText">&gt; cgroups are needed, but why would you disable them?</font><br> e.g. I see no reason for them (and the system in question isn't the one where stray processes are welcome in the first place).<br> <p> <font class="QuotedText">&gt; There's also nothing wrong with systemd's use of D-Bus.</font><br> ...but <a rel="nofollow" href="http://secunia.com/advisories/search/?search=dbus">http://secunia.com/advisories/search/?search=dbus</a><br> <p> <font class="QuotedText">&gt; And besides, systemd is by no means the only program that uses cgroups.</font><br> Do you know the difference between PID 1 and a bunch of other processes, Mr. Downright Downplayer?<br> <p> <font class="QuotedText">&gt; thus making systemd portable would be a waste of time.</font><br> What about accepting patches?<br> <p> <font class="QuotedText">&gt;&gt; modularity.</font><br> <font class="QuotedText">&gt; its configure script...</font><br> Next time you buy bread hope they won't suggest you a grocery shop instead; but if they accidentally do, please make a mental note that it isn't you asked for.<br> <p> What is irritating in the crowd touting systemd is that viral "who would need this or that?". You never know in advance. Just ask Linus if you haven't read the story already: did he anticipate an Alpha port, or Beowulf clusters, or even thousands of users at the outset? And yet this egocentric upstream -- which is churning out a lot of code indeed -- damaged their heads enough so as to think that they can decide for the rest of us.<br> <p> They can't.<br> <p> Heck, I've suggested this article in the distro's devel mailing list just in case OpenRC might get considered and found useful -- we thoroughly dislike an idea of running systemd at servers.<br> </div> Thu, 30 Aug 2012 18:04:29 +0000 Debian looks at OpenRC https://lwn.net/Articles/513818/ https://lwn.net/Articles/513818/ HelloWorld <div class="FormattedComment"> <font class="QuotedText">&gt; In particular, if anyone could demonstrate a way of doing this in a »minimally invasive« manner,</font><br> It can't be done. This email mentions an incomplete list of non-portable interfaces that systemd uses. <br> <a href="http://lists.debian.org/debian-devel/2011/07/msg00281.html">http://lists.debian.org/debian-devel/2011/07/msg00281.html</a><br> The only way to make systemd portable is to essentially rewrite it.<br> </div> Tue, 28 Aug 2012 13:21:39 +0000 Debian looks at OpenRC https://lwn.net/Articles/513620/ https://lwn.net/Articles/513620/ anselm <blockquote> <em>They wouldn't be where they are today if portability was made a goal at the project outset. I still hope that one of the BSDs pick it up and make their own portable version because I think the systemd approach is superior to the old systems.</em></blockquote> <p> In particular, if anyone could demonstrate a way of doing this in a »minimally invasive« manner, the systemd developers might be brought around to supporting the required changes when the code has otherwise settled down. It makes a lot of sense to produce a Linux version of systemd first and then see what, if anything, can be done about portability. </p> <p> On the other hand, it would mean somebody with a stake in BSD would have to do it, and that is unlikely to happen because of NIH. </p> Mon, 27 Aug 2012 10:22:40 +0000 Debian looks at OpenRC https://lwn.net/Articles/513533/ https://lwn.net/Articles/513533/ HelloWorld <div class="FormattedComment"> <font class="QuotedText">&gt; The tight-coupling between components which systemd is encouraging and requiring is something which, for those who remember back then, we used to criticise Microsoft for, due to their inability to dig themselves out of the rut such philosophy had led to, being unable to fix even trivial bugs due to fixing them breaking old code. It leads to inflexible systems which can't change, whereas loosely-coupled systems, which Linux has had until now, allow components to be easily swapped out and changed providing that the interfaces between the components are relatively well specified.</font><br> Systemd's interfaces are perfectly well-specified.<br> <a href="http://www.freedesktop.org/wiki/Software/systemd/InterfaceStabilityPromise">http://www.freedesktop.org/wiki/Software/systemd/Interfac...</a><br> They're also relatively simple and portable, so that other init-like software can use the same interfaces for e. g. socket activation. I'm sorry, but I don't see your point at all.<br> <p> <font class="QuotedText">&gt; The loose coupling does have the disadvantage that components aren't necessarily as well integrated, but this hasn't really been a problem in practice--the benefits far outweigh these downsides,</font><br> That's an easy claim to make given that you provide no arguments to back it up and it's thus essentially impossible to refute. I can't help noticing though that the successful desktop operating systems do apply a well-integrated approach.<br> <p> <font class="QuotedText">&gt; An example is the requirement for stuff like cgroups, autofs, D-Bus etc. Having these as a requirement, rather than merely being optional and used if available, imposes constraints on other projects. cgroups can never be removed from the kernel, and distribution kernels must include it.</font><br> Systemd doesn't strictly require autofs. cgroups are needed, but why would you disable them? The overhead is negligible if you disable all cgroup controllers, none of which are required for systemd. There's also nothing wrong with systemd's use of D-Bus. It doesn't require running dbus-daemon, and there's nothing wrong with the protocol; in fact, it would have been brain-dead to use anything else, we don't need yet more NIH syndrome on the linux desktop. <br> <p> <font class="QuotedText">&gt; So it indirectly forces other parties not to allow certain components to evolve or be removed.</font><br> By that logic, every kind of code reuse is bad because it forces the code that is being reused to keep a stable interface. I'm sorry, but that's downright bullshit. If the cgroup stuff is actually so bad that the kernel folks want to remove it again, it shouldn't have gone in in the first place. They've made their bed, now they must lie in it. And besides, systemd is by no means the only program that uses cgroups. <br> <p> <font class="QuotedText">&gt; If systemd was a bit less aggressive in its strict requirements, it would be rather more portable</font><br> Every Unix-like OS ships its own init system: SMF, launchd, BSD init etc.. Do you actually expect any of those to give up their init system in favour of systemd? I certainly don't, thus making systemd portable would be a waste of time.<br> <p> <font class="QuotedText">&gt; Systemd certainly has a lot of good ideas, but the package as a whole is does not allow one to pick and choose the good from the bad due to its lack of modularity.</font><br> This is yet more nonsense, systemd is quite configurable. Just look at its configure script...<br> </div> Sun, 26 Aug 2012 00:17:17 +0000 Debian looks at OpenRC https://lwn.net/Articles/513508/ https://lwn.net/Articles/513508/ raven667 <div class="FormattedComment"> <font class="QuotedText">&gt; Klumpp said: "So far nobody has created a non-Linux fork of systemd, and the reason is mainly that it is too much work."</font><br> <p> Which makes it clear why the systemd team isn't jumping up, dropping everything and sacrificing all that time on portability. They wouldn't be where they are today if portability was made a goal at the project outset. I still hope that one of the BSDs pick it up and make their own portable version because I think the systemd approach is superior to the old systems. BSD projects like OpenSSH show the way it can be done. <br> </div> Sat, 25 Aug 2012 17:16:55 +0000 Debian looks at OpenRC https://lwn.net/Articles/513323/ https://lwn.net/Articles/513323/ HelloWorld <div class="FormattedComment"> <font class="QuotedText">&gt; Why?</font><br> Because every moving part in the system increases the amount of testing that needs to be done in order to ensure that things actually work together well. Of couse, that kind of testing never happens, so what we get isn't choice between several different, but working implementations, but choice between several different kinds of crap, the least bad of which one has to put up with. <br> <p> <font class="QuotedText">&gt; Says who? Different applications and workloads may need different init systems. A person may need a "give me a text shell as fast as you can" or "restart the database really fast" or "show me my graphical desktop, even before the network is up" use cases. Some workloads are better served by upstart's logging and start-daemon-on-demand. There is no init system whose size fits all.</font><br> Name one inherent reason a single init system can't do all of those things well.<br> </div> Fri, 24 Aug 2012 14:16:43 +0000 Debian looks at OpenRC https://lwn.net/Articles/513302/ https://lwn.net/Articles/513302/ mpr22 The construction "X of the week" tends to bring to mind plot devices popular in episodic TV e.g. the "monster of the week" in an action show (shows up at start of episode, wreaks havoc, gets beaten by the heroes by the time the end credits roll) or the local Casanova's "girlfriend of the week" (shows up at start of episode, instantiates some allegedly-amusing personality stereotype, dumps him by the time the end credits roll) in a sitcom. Calling something that has been around in wide use for ten years "X of the week" seems a suboptimal use of the term. Fri, 24 Aug 2012 11:22:37 +0000 Debian looks at OpenRC https://lwn.net/Articles/513272/ https://lwn.net/Articles/513272/ dgm <div class="FormattedComment"> A highly coupled system is one where any part of it can interact with any other part of it, like a program written in assembly language, where every line of code is a potential jump destination. This makes changing parts rather difficult, as stated.<br> <p> A loosely coupled one is just the opposite. In those all interactions happen through interfaces, so as long as the interfaces are respected, you can confidently change the parts behind them. <br> </div> Fri, 24 Aug 2012 07:06:15 +0000 Debian looks at OpenRC https://lwn.net/Articles/513253/ https://lwn.net/Articles/513253/ walters <div class="FormattedComment"> systemd only depends on libdbus, not the message bus daemon. There's quite a bit of difference between the two.<br> <p> <p> </div> Fri, 24 Aug 2012 01:57:40 +0000 Debian looks at OpenRC https://lwn.net/Articles/513235/ https://lwn.net/Articles/513235/ viro <div class="FormattedComment"> OK, OK - "coprolite of the week". Pedants... (Unless I've misparsed you and you mean that no piece of software remains shitty ten years after it had been started, in which case I have a really nice bridge for you...)<br> </div> Thu, 23 Aug 2012 22:41:25 +0000 Debian looks at OpenRC https://lwn.net/Articles/513233/ https://lwn.net/Articles/513233/ daniels <div class="FormattedComment"> <font class="QuotedText">&gt; It leads to inflexible systems which can't change, whereas loosely-coupled systems, which Linux has had until now, allow components to be easily swapped out and changed providing that the interfaces between the components are relatively well specified.</font><br> <p> Wait, what? If you have a single large and totally integrated system, you can change what you want: everything is an implementation detail. If you have an infinitely modular system, you can't change anything, because every line of code is an interface.<br> <p> I don't see your argument at all.<br> </div> Thu, 23 Aug 2012 22:26:20 +0000 Debian looks at OpenRC https://lwn.net/Articles/513231/ https://lwn.net/Articles/513231/ daniels <div class="FormattedComment"> <font class="QuotedText">&gt; 1) take a turd-of-the-week (dbus, etc.)</font><br> <p> 'Of the week'? It's ten years old!<br> </div> Thu, 23 Aug 2012 22:23:41 +0000 Debian looks at OpenRC https://lwn.net/Articles/513224/ https://lwn.net/Articles/513224/ rleigh This is one of my primary objections to systemd, which was mentioned here: <a href="http://lists.debian.org/debian-devel/2012/04/msg00751.html">http://lists.debian.org/debian-devel/2012/04/msg00751.html</a>. <p> The <b>tight-coupling</b> between components which systemd is encouraging and requiring is something which, for those who remember back then, we used to criticise Microsoft for, due to their inability to dig themselves out of the rut such philosophy had led to, being unable to fix even trivial bugs due to fixing them breaking old code. It leads to inflexible systems which can't change, whereas loosely-coupled systems, which Linux has had until now, allow components to be easily swapped out and changed providing that the interfaces between the components are relatively well specified. The loose coupling does have the disadvantage that components aren't necessarily as well integrated, but this hasn't really been a problem in practice--the benefits far outweigh these downsides, and this seems to have been forgotten in some camps. <p> An example is the requirement for stuff like cgroups, autofs, D-Bus etc. Having these as a requirement, rather than merely being optional and used if available, imposes constraints on other projects. cgroups can never be removed from the kernel, and distribution kernels <b>must</b> include it. So it indirectly forces other parties not to allow certain components to evolve or be removed. Likewise in userspace with D-Bus. I can't say I'm particularly happy to have such a nasty bit of code required to be running on my system, let alone being used by init. If systemd was a bit less aggressive in its strict requirements, it would be rather more portable (including to Linux, for kernels which don't have all the extra bits built in). <p> The other issue is the reliability of the system. sysvinit is tiny (~30KiB on i386), because it does relatively little. This makes it very reliable. PID 1 must never die! There's no <i>need</i> for PID1 to do anything but the bare minimum--it's quite fine for more advanced stuff to be done by other processes forked from init. The ideas behind other less well known alternatives such as <a href="http://www.skarnet.org/software/s6/index.html">s6</a> are also well worth looking into. The nice thing about OpenRC is that it's run <i>from</i> init, so it can do some more advanced stuff like dependency-based boot, but the extra complexity is not going into PID 1. Now, it's clearly not as advanced as systemd. But it's most certainly more advanced than sysv-rc. It gives us <i>some</i> of the features that systemd offers, but without the massive downsides. <p> Lastly, while there's a lot of momentum in the adoption of systemd across several other distributions, I always feel somewhat bothered by the arguments that we must adopt it or get left behind. Blindly jumping on the latest bandwagon is rarely a sensible choice, particuarly when it might not be taking us where we want to go. And while having a single init across distributions is touted as a generally good idea, having different implementations does prevent gross stupidities from taking place, nor give any single party the ability to force unwanted changes upon eveyone. Systemd certainly has a lot of good ideas, but the package as a whole is does not allow one to pick and choose the good from the bad due to its lack of modularity. That might require them to be reimplemented. As has happened with cgroups in OpenRC. Thu, 23 Aug 2012 21:50:20 +0000 Debian looks at OpenRC https://lwn.net/Articles/513218/ https://lwn.net/Articles/513218/ hummassa <div class="FormattedComment"> <font class="QuotedText">&gt; That sounds terrible. </font><br> <p> Why?<br> <p> <font class="QuotedText">&gt; What Debian really needs to do is just make a choice. They only need one init system.</font><br> <p> Says who? Different applications and workloads may need different init systems. A person may need a "give me a text shell as fast as you can" or "restart the database really fast" or "show me my graphical desktop, even before the network is up" use cases. Some workloads are better served by upstart's logging and start-daemon-on-demand. There is no init system whose size fits all.<br> </div> Thu, 23 Aug 2012 20:42:54 +0000 Debian looks at OpenRC https://lwn.net/Articles/513188/ https://lwn.net/Articles/513188/ viro <div class="FormattedComment"> ... because all this fragmentation makes it so much harder to play dependency games, doesn't it?<br> <p> Here's how it works:<br> 1) take a turd-of-the-week (dbus, etc.)<br> 2) make systemd depend on it<br> 3) push patches adding dependencies on said turd to as many packages as you can, solidifying the damn thing to the point where it would be very hard to avoid even on the systems that are not systemd-infested.<br> 4) pick the next turd, repeat the whole thing.<br> <p> With systemd on every box the whole thing would become so much easier - just cut (3) out, avoiding the need to convince a lot of developers and (oh, horror) to deal with their objections.<br> </div> Thu, 23 Aug 2012 18:57:15 +0000 Debian looks at OpenRC https://lwn.net/Articles/513151/ https://lwn.net/Articles/513151/ josh <div class="FormattedComment"> I agree entirely, but at least with a compatibility translator like this, the systemd .service files could become the canonical description of how to run a service, and packagers would not need to write sysvinit scripts by hand anymore. Seems like the easiest way to "support" sysvinit without forcing every service to deal with sysvinit scripts, which seems like a net improvement over the current state.<br> </div> Thu, 23 Aug 2012 16:56:19 +0000 Debian looks at OpenRC https://lwn.net/Articles/513143/ https://lwn.net/Articles/513143/ drag <div class="FormattedComment"> That sounds terrible. <br> <p> What Debian really needs to do is just make a choice. They only need one init system.<br> </div> Thu, 23 Aug 2012 16:22:28 +0000 Debian looks at OpenRC https://lwn.net/Articles/513140/ https://lwn.net/Articles/513140/ kaol Yet another option would be to automatically generate init scripts for various init systems from a single source. This year's GSoC <a href="http://wiki.debian.org/SummerOfCode2012/StudentApplications/akhilvij">had one such project</a>, though it yet remains to be seen if and how it will be used in Debian. Akhil's project used systemd init files as the source, which seems like a good choice to me, given how declarative they are. <p>LSB init scripts are full of boilerplate. Thu, 23 Aug 2012 16:11:39 +0000 Debian looks at OpenRC https://lwn.net/Articles/513019/ https://lwn.net/Articles/513019/ dlang <div class="FormattedComment"> If we have a good dependency based init, it seems like it would not be that big a stretch to add a dependency on some state, and then let whatever other process is appropriate set that state.<br> <p> This could be as simple as waiting for a special file to be created, or for a message to be written to a known named pipe.<br> </div> Thu, 23 Aug 2012 03:19:13 +0000