Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Debian To Replace SysVinit, Switch To Systemd Or Upstart 362

An anonymous reader writes "Debian has been one of the last holdouts using SysVinit over a modern init system, but now after much discussion amongst Debian developers, they are deciding whether to support systemd or Upstart as their default init system. The Debian technical committee has been asked to vote on which init system to use, which could swing in favor of using Upstart due to the Canonical bias present on the committee."
This discussion has been archived. No new comments can be posted.

Debian To Replace SysVinit, Switch To Systemd Or Upstart

Comments Filter:
  • by GameboyRMH ( 1153867 ) <gameboyrmh@NoSpam.gmail.com> on Monday October 28, 2013 @01:41PM (#45260513) Journal

    But that doesn't mean that Upstart does.

    • Re: (Score:3, Interesting)

      by Anonymous Coward

      Yea who doesn't want their DB servers kill -9'd if they take too long to shutdown. Hooray upstart.

    • Re: (Score:2, Interesting)

      by Anonymous Coward

      IMHO, it does (suck) when compared to Systemd.

      I've also found moving from System5 based init's easier with Systemd

    • Upstart sucks.

      Every time I've tried to make new software be controlled by it, I've wound up writing horrendous kludges or (where possible) rewriting the source code to make it more upstart-friendly.

      Upstart is to init systems what Unity is to desktops - mostly okay for some stuff, and utter **** for anything else.

    • by fatphil ( 181876 ) on Monday October 28, 2013 @03:06PM (#45261539) Homepage
      Indeed, Canonical sucking *doesn't* mean that Upstart sucks.
      It's the fact that Upstart sucks which means that Upstart sucks.

      At least from an embedded perspective (which is the majority of linux systems) the system should start only that which is necessary rather than everything that is possible. It also suffers from the classic stampede mistake when it becomes possible to start a whole load of new services after a shared dependency is started.
  • Ugh (Score:5, Insightful)

    by ArchieBunker ( 132337 ) on Monday October 28, 2013 @01:50PM (#45260611) Homepage

    I fucking hate this new system. Its a mess of scripts that call on more scripts. Its such a pain in the ass now if you want to have a program run when the system starts. Gone are the days of just adding a line to /etc/rc.local

    • Re:Ugh (Score:5, Interesting)

      by lasermike026 ( 528051 ) on Monday October 28, 2013 @01:52PM (#45260635)

      Yeah, SysVinit is not only fine but preferable. It's simple and effective. I see no reason to change.

      • Re: (Score:3, Funny)

        I'm with this too. Fuck Canonical. They have become a real pain in the ass, and worst of all, they're a pack of fucking retarded assholes. I can't say it enough. Fuck Canonical. Fuck Canonical. Fuck Canonical.

        Debian is one of the best distros out there, let's not allow the fuckwaddery that is Canonical and their arrogance and intense stupidity ruin it. The current init system works just fine.

        Oh, and in case you didn't get it. Fuck Canonical. Fuck Canonical. Fuck Canonical.

        • Re:Ugh (Score:5, Funny)

          by Tanktalus ( 794810 ) on Monday October 28, 2013 @02:00PM (#45260707) Journal

          I really wish you'd get off the fence and pick a position on the issue.

        • Really? (Score:4, Interesting)

          by prefec2 ( 875483 ) on Monday October 28, 2013 @02:12PM (#45260855)

          This is your way of stating your opinion? Looks, like your parents forgot to teach you something particular important to be considered an adult and able to participate in a discussion. You could say that you disagree with Canonicals decision towards Wayland/Weston. You could say that the way they handled the issue and announced Mir sucks and was not a cooperative mode. In addition they made some bad blood. And that all would sound like someone really discussing the issue, but really you just sound like someone either too young to remember the vi vs. emacs wars or like someone exhumed from that day, just to behave like a troll.

          BTW: Get a life, normally that reduced these urges to hate someone just because you find his or her decision questionable.

          • Re:Really? (Score:4, Interesting)

            by jedidiah ( 1196 ) on Monday October 28, 2013 @02:19PM (#45260955) Homepage

            It's far too easy to shoot yourself in the foot with upstart. It is much more complicated. In it's efforts to solve a non-problem (namely making a machine boot faster) it tries to do things in parallel and creates a web of dependencies that can render your system unbootable.

            It adds extra complexity for dubious gain.

            It's the perfect example of why we shouldn't necessarily accept the advice of "helpful types" that think that things should be done the way that Microsoft does it or the way Apple does it.

            • by coder111 ( 912060 ) <{moc.liamrr} {ta} {redoc}> on Monday October 28, 2013 @04:08PM (#45262209)
              I do agree bootup times don't matter if you run a server. For a laptop, a tablet, a mobile, even a desktop that gets turned off startup times are important. For tablets, laptops and mobiles, they are VERY important.

              I agree that complexity is evil. I have no experience with systemd nor with upstart, so I cannot comment on them. However, dependency graph and parallel execution should not be THAT difficult or complex :-/

              • Re: (Score:3, Interesting)

                by Anonymous Coward

                Yes, complexity is evil, the problem is that sysv is complex (read: bash is complex) and running bash's init code hundreds (thousands?) of times and seeking all over disk eats lots of time.

                The solution therefore, is not to introduce something even uglier and more complex like Upstart or systemd, which STILL runs bash init code hundreds of times, and seeks all over disk, but now also adds megabytes of source code that one either has to audit or put blind faith in, even though both systems are empirically unr

            • It *shouldn't* matter how long it takes your server to boot, but it really could be critical in the event of an outage. In any case systemD solves more problems than optimising boot times.

          • To me it seems the Slashdot hive mind is starting to teach everyone to robotically hate Canonical.
    • Re:Ugh (Score:5, Informative)

      by Arker ( 91948 ) on Monday October 28, 2013 @01:59PM (#45260695) Homepage
      Slackware still uses sensible init scripts you know.
      • At a former university the desktops used Slackware with initscripts all the way up until 2011. These desktops were beefy machines with decent chunks of RAM and SATA disks but would take minutes to finish booting. The administrators there switched the start-up scripts to systemd across a holiday and things became dramatically better - no boot took more than 30 seconds and most disk based boots were sub 15 seconds making the old system look laughable.

        I'm glad that Slackware gave the admins the choice because

        • by Arker ( 91948 )

          I havent noticed any such slowness with slackware startup myself, so I suspect a misconfiguration, but nonetheless, even if that were the cost, having sensible human-readable startup scripts is well worth a couple extra minutes boot time to me anyway. Seriously, how often do you reboot a *nix box? No one really cares how long it takes to do that one-in-a-blue-moon reboot.

    • Re:Uh... (Score:5, Informative)

      by Chemisor ( 97276 ) on Monday October 28, 2013 @02:04PM (#45260757)

      I fucking hate this new system. Its a mess of scripts that call on more scripts.

      Actually, that's how sysv init works. To get a program started by systemd you have to create a service file full of magic commands and put it in the magic systemd directory. Then you have to type systemctl --abracadabra enable yourservicename.service. Then you have to go and add an [install] section to your service file, because nobody actually remembers that you have to write one or how to do it. Then you do the systemctl again. Then you check the log files to see if the thing actually started, because nothing gets output to the console during boot (except the filesystem mount messages and the big fat warning that my root fs is readonly).

    • Re:Ugh (Score:5, Insightful)

      by icebike ( 68054 ) on Monday October 28, 2013 @03:17PM (#45261643)

      I fucking hate this new system. Its a mess of scripts that call on more scripts. Its such a pain in the ass now if you want to have a program run when the system starts. Gone are the days of just adding a line to /etc/rc.local

      Half of that is because either SystemD or upstart is really only about half implemented, and the half that is implemented is often trying to replicate sysv just to keep the conversion and learning task to something approaching manageable. Its kind of a mess right now in many distros.

      As more of the system targets are properly implemented, and users start to let go of the concept of run levels, and get used to dealing with target files and the concept of units, it will be every bit as tailor-able as run levels were, and a whole lot faster.

      I didn't find run levels and rc.d all that intuitive at first (many long years ago) and the scripts were more complex.

    • Re:Ugh (Score:4, Informative)

      by thrift24 ( 683443 ) on Monday October 28, 2013 @04:02PM (#45262129) Homepage
      It sounds like you are describing the current system, not the new system. And if you don't like the current system, you are going to hate the new system.

      What I would consider the old system would be something like SlackWare(RC scripts) where the system might have 8 or 9 different shell scripts called during the boot process, but it's essentially one giant autoexec.bat. Simple to make modifications to the boot process, but installing new services or upgrading the system may require manual merges and break your installation.

      The current system is more like RedHat/SuSE/debian(Tradtitional init), where there are tons of scripts that call other scripts and it gets pretty complicated, but for the most part everything is a script and can be easily traced. This is more difficult for the inexperienced to modify, but is reasonable for those familiar with scripting and is great for adding new services and upgrading the OS. Basically the scope of a change is smaller, so less stuff breaks.

      The new system is something like Fedora/Ubuntu(Systemd/Upstart). There are config files for everything from services, to devices, to sockets that are parsed by a binary that isn't very open to inspection. This leads to a very fast boot up and has neat features like the ability to view the logs of a service with the same command used to start it, but is also like sticking your dick in a box of razers, because when something goes wrong you can't just pull out vi and look at the logic being used to boot the machine. It also leads to somewhat unsettling things like a merged /usr and a /run.

      To be fair this might be somewhat unfair to Ubuntu, because I haven't invested much time into Upstart. If it was something worth looking at I'm sure the Fedora/SUSE devs would have dropped systemd for it though.

      What would be really nice is if Debian built a version of systemd that didn't have a big binary core, or at least split the thing into several different services. I like the speed and slickness of systemd, but if anything goes wrong with a system using it, I will have no idea how to fix the thing -- and that's after using systemd on my primary laptop/server for over a year.
  • by Arker ( 91948 ) on Monday October 28, 2013 @01:57PM (#45260673) Homepage
    Init [slackware.com] would have been my pick, but I still hope this works out well for them.
  • How about OpenRC? (Score:4, Interesting)

    by Anonymous Coward on Monday October 28, 2013 @02:01PM (#45260721)

    Why not keep sysvinit and switch to OpenRC for managing the init scripts?

  • by helobugz ( 2849599 ) on Monday October 28, 2013 @02:06PM (#45260789)

    Tell me I'm not the only one still clinging to sysvinit?

    The new "replacements" (alternatives) are ghey++ and will no doubt be replaced in due time.

    I dno't want to hear about a few seconds faster boot time. I want my *nix startup to be configurable, scripted, and simple. sysvinit takes the cake;.It's documented, and sysvinit is so simple it doesn't really require documentation, anyway.

    • by Anonymous Coward on Monday October 28, 2013 @02:17PM (#45260937)

      On a Gentoo box, and it should still be starting via sysvinit.

      My #1 reason for keeping it installed is standardization:
      All the BSDs use a similiar system, all the legacy UNIXes do as well, as do all my linux installs that are more than 2 years old.

      Additionally I have had *NO* problems with it in like 15 years that weren't caused by user error, or distro error. Systemd on the other hand rendered my system hung or inoperable on more than a few occasions when it first became popular, as has udev by itself. There's something to be said for 'windows-like' functionality, but all the subsystems that have been getting added to linux to provide it are proving messy, unmaintainable, and even more prone to 'unidirectional grading' (it used to be you could have both newer and older kernels, even across major versions running. Nowadays you're lucky if the minor versions don't break things over the span of two months. Anyone here remember having 1.2 installs running 2.0? Or 2.0 with a 2.2 kernel? Or 2.2/2.4? The only major issues you had were if you used ipchains/tables/ipfwadm and had to migrate your settings. And there was almost always legacy support for most or all of a major version change.)

      Honestly with the way linux is going nowadays, as well as the various *BSDs, I'm considering very strongly migrating to another platform. If you change what people are used to too much, there's far far FAR less incentive for them not to try something totally new rather than bungling themselves up with half remembered details about how their *FORMER* version of the system operates. Much like happened with WinXP/Vista/7/8.)

      Not that many people will agree with this assessment.

      • by 0123456 ( 636235 )

        Not that many people will agree with this assessment.

        No, I've been thinking that too. The more they try to make Linux like Windows, the less reason there is not to consider other operating systems.

        It's certainly true that making startup work reliably with init scripts can be hard: service foo has to start after service bar, which can only start after the network is up, or whatever. But none of the alternatives seem appealing yet.

        • Re: (Score:3, Informative)

          by hson ( 78256 )
          IMHO NetBSD's rc-system works really well here (Used by FreeBSD nowadays too).
          Simple and effective - and no symlinks all over the place like in SysV. It automagically orders services. If foo needs bar bar is started before foo.
          And I see now that it is what OpenRC is... or atleast was.
    • Opensuse dropped sysvinit when the current level came out several months ago. Before that they had defaulted to systemd but allowed sysvinit.
      My considered opinion of systemd? I hate it. I don't have a server farm, I have a couple of desktop machines with Samba and nfs capability and sysvinit was ideal.

    • I still use SysV because solutions are portable. I work on every flavour of UNIX and some I shouldn't even be able to run anyway. I don't want to write solutions for multiple inits.

      SysV for as long as they'll let me. My systems will stay up longer, take less work to maintain and configurations be clearly apparent to actual sysadmins.

      Just my two cents.

  • by Anonymous Coward on Monday October 28, 2013 @02:14PM (#45260897)

    I used to not only use RedHat, but contributed fixes. Ubuntu is not, in many respects, my ideal, but systemd is enough reason all by itself for me to use Ubuntu.

  • by sl4shd0rk ( 755837 ) on Monday October 28, 2013 @02:16PM (#45260923)

    SysV init scripts just work. They are simple to create and easy to maintain. Debugging is a cinch when things go wrong -- and a lot of packages DO NOT get upstart init, nor SysV init correct. Upstart is only easier in theory. In practice it's made a complete mess of things and I have several Ubuntu systems to prove it.

    • by Rich0 ( 548339 ) on Monday October 28, 2013 @02:23PM (#45261013) Homepage

      The problem with sysvinit is that 95% of daemons just need a common set of actions to start/stop them, and sysvinit tend to handle this by writing bash scripts starting from a skeleton.

      So maybe for one daemon you can set a config setting to make it use ionice, and for another you can't, simply because in one script a bit of extra functionality was written.

      For the most part systemd makes a config file into a config file, not an executable. Sure, you can run a script if you have to, but 99% of the time you don't need to.

      In fact, there is no reason you couldn't write a sysvinit script that takes in a systemd unit as a config file and starts/stops the corresponding service. That would be a great way to transition - switch your bash scripts to unit files gradually and then swap out the init system.

    • by jedidiah ( 1196 )

      Anyone trying to tell you that Upstart is easier is simply lying to you.

      Upstart is more complex by design.

    • by whois ( 27479 ) on Monday October 28, 2013 @02:59PM (#45261465) Homepage

      Debian has a habit of not using things until they work. I expect they would fix most of the issues or they wouldn't ship it.

  • This sounds like a really awful idea, based on what I've read here. I like how Debian's init works now, its fine, there's nothing wrong with and it's simple.

    Where the heck do I send hatemail to perhaps encourage them not to do this?

  • by shallot ( 172865 ) on Monday October 28, 2013 @02:28PM (#45261075)
    Citation needed, anonymous. When has the Debian Technical Committee last made a decision based on a political bias?
  • Finally! (Score:5, Interesting)

    by cpicon92 ( 1157705 ) <kristianpicon@gmail.com> on Monday October 28, 2013 @02:29PM (#45261077)

    It seems like I'm the only person on here who thinks this, but I really can't wait for the switch to happen. Upstart scripts are unbelievably easy to write when compared with init scripts. For one thing, they don't require massive amounts of boilerplate code. For another, they are relatively easy to manage and execute.

    Just the other day I was trying to set up a couple of machines running deluge as a daemon. The Ubuntu machines took me 10 minutes tops. The remaining debian one had me in init script hell for an hour or more...

    • Massive amounts of boilerplate code?

      I guess we've been using different SVR4 variants. I just use a case statement to test the first argument...

      My beef with upstart et al is that they disable, instead of deprecating, old init. That means now I have to maintain at least three sets of scripts instead of two (sysv, bsd) for software distribution.

      And what the hell was wrong with inittab? Let's see, it's succinct, works, and is fully debugged.

  • Upstart (Score:4, Insightful)

    by intangible ( 252848 ) on Monday October 28, 2013 @02:33PM (#45261133) Homepage

    So upstart has some things that need to be fixed (mostly the clean shutdown thing)...
    Systemd is a monster that gets to infect more of you packages over time, plus you get the benefit of binary log files!

    I hope they choose upstart and just fix it up a bit.

    OpenRC has been proposed by some too, which seems like a nice sysvinit replacement, but event driven startup and shutdown of services (think laptops and hotswap stuff) is more important than just a fast startup time.

    • by gmueckl ( 950314 )

      I honestly wonder what the systemd developers were thinking when they turned it into a feature-creep laden mountain of mostly annoying features which slowly takes over the system from you. The way it seems to force itself into other things in the system (e.g. by way of systemd-specific modifications in daemons and such) just should have set off a lot of software engineering alarm bells. Why didn't that happen?

  • None too soon! (Score:4, Interesting)

    by redelm ( 54142 ) on Monday October 28, 2013 @02:42PM (#45261255) Homepage
    I cannot abide the SysV (AT&T) mess'o'symlinks multiple-indirect startup scripts. One reason I've stuck with Slackware for almost 20 years. It uses BSD-style inits that have far less indirection and need far fewer lookups. Frankly, some of the BusyBox startup look attractive too -- one script to rule them all :)
  • by Anonymous Coward

    Debian should stay true to its core principles. The user should be able - by installing a package - to choose the init system she wants.
    There is no one true mail server on Debian. You can choose to run zsh as a shell instead of bash. There should never be only one supported init system.
    Distributions don't choose, users do. Oh and technical committees are there to find solutions to make that work, not to complain about how difficult that is.

  • Hoping for systemd (Score:5, Informative)

    by devent ( 1627873 ) on Monday October 28, 2013 @03:02PM (#45261493) Homepage

    I hope for systemd; I know it from Fedora. And in my opinion upstart is some kind of mess; it's a mixture of bash script and their own added syntax. It kind of feels like Microsoft's extensions for C++. I'm also a fan of declarative configuration like systemd is. After 5 minutes reading the manual of systemd I could write my own service for pdnsd.


    ExecStart=/usr/local/sbin/pdnsd --daemon -p /var/run/pdnsd.pid


    # systemctl status pdnsd
    pdnsd.service - PDNSD
          Loaded: loaded (/usr/lib/systemd/user/pdnsd.service)
          Active: active (running) since Mon 2013-10-28 18:46:23 CET; 1h 14min ago
        Process: 1585 ExecStart=/usr/local/sbin/pdnsd --daemon -p /var/run/pdnsd.pid (code=exited, status=0/SUCCESS)
      Main PID: 1587 (pdnsd)
          CGroup: name=systemd:/system/pdnsd.service
                          1587 /usr/local/sbin/pdnsd --daemon -p /var/run/pdnsd.pid

    Oct 28 18:46:23 vostrotitan.localdomain systemd[1]: Starting PDNSD...
    Oct 28 18:46:23 vostrotitan.localdomain pdnsd[1587]: pdnsd-1.2.9a-par starting.
    Oct 28 18:46:23 vostrotitan.localdomain systemd[1]: Started PDNSD.

    • by dshk ( 838175 ) on Monday October 28, 2013 @06:42PM (#45263655)
      Here is an Upstart script copied directly from the minecraft server of my 11-year-old son:

      start on runlevel [2345]
      stop on runlevel [^2345]

      setuid attila
      setgid mc
      env LANG=en_US.UTF-8
      chdir /home/attila/minecraft

      exec java -client -Xms2500M -Xmx2510M -jar spigot.jar > /dev/null

      And then:
      # start minecraft

      I cannot see a single line which requires documentation in order to understand it.

  • by Anonymous Coward


    • by Aguazul2 ( 2591049 ) on Monday October 28, 2013 @04:00PM (#45262115)

      Wow, with such a superior, arrogant and manipulative attitude, I think it is time for Debian to write their own, or continue with what they have. I would have nothing to do with someone who has so little respect for anyone else's efforts, using FUD against other projects, and who is so obviously trying to lure people into his self-serving spider-web trap. Perhaps his vision is that systemd will be the one process to rule them all, my precioussssss... and then finally Linux will be all his. Run away!!!

  • by Anonymous Coward


    1. What sane organization wants to follow in Canonical's footsteps?
    2. Who, besides an utter fool, thinks systemd is a good idea?

  • FTS: The Debian technical committee has been asked to vote on which init system to use, which could swing in favor of using Upstart due to the Canonical bias present on the committee."

    So what are the chances of getting the Canonical-associated board members to recuse themselves from the vote, given the obvious conflict of interest there?

  • by waffle zero ( 322430 ) on Monday October 28, 2013 @04:44PM (#45262539) Journal

    And that is a good thing for Linux because it can use a lot of good technology from the kernel. The major issue is that systemd requires cgroups and that means no support for kFreeBSD. Even if the ex-Canonical people recused themselves, systemd was always going to have an uphill battle.

    There is a Debian derivative that has decided to use systemd, but it's -- the still incubating -- Tanglu.

  • I'm Torn (Score:5, Interesting)

    by Foresto ( 127767 ) on Monday October 28, 2013 @04:46PM (#45262549) Homepage

    After having repeatedly run into the limitations of SysV init, I'm all for replacing it with something smarter, but I'm torn between these two.

    I've used Upstart on Ubuntu, both as an admin and as a developer. I like that the commands and configuration files are clean and pretty easy to understand. A few things bother me, though:

    • The model of starting all dependent services when their dependencies start is backwards. I don't necessarily want init to launch every daemon under the sun the moment I mount their data filesystem. I'd rather have it mount the required filesystem when I ask for a particular daemon to start.
    • As of a year or so ago, the documentation was mainly an incomplete bunch of blog posts. Once I found them, it was pretty easy to configure daemons that behaved like the venerable ones that are often used as examples, but it was difficult to learn how to match Upstart's features (some of which are undocumented) and events (also largely undocumented) with an unusual service's behavior
    • Debugging was difficult, mainly because so few events are well documented and it's not always clear which of Upstart's features are implemented in in any given version. (I hear the latest release offers some event tracing tools that would improve this.)

    I haven't used Systemd at all, but the common points that come up again and again in every writeup I encounter have me forming me some opinions already. I really like the idea of the load-as-needed dependency model. A few things have me quite worried about the implementation, though:

    • Configuration is reportedly difficult to understand. That always leads to frustrating, time-wasting, messy problems.
    • The code is reportly rather complex. That usually leads to chronically buggy software, which is not what I want in a process as important as init. It also tends to hamper portability, which could make Systemd a poor candidate for replacing init on other unixes, which would relegate it to being yet another OS-specific hassle for coders and admins all over the world. I'd prefer something that could reasonably be adopted everywhere, or at least by most of the operating systems I have to administer, even if a few features weren't available on every platform
    • I recently learned that the guy behind Systemd is the same guy who brought us PulseAudio. I don't want to get off topic here, but this gives me little hope that Systemd will ever work well outside the lead developers' development machines. (Pulse is around 10 years old now, and every time I give it another chance, it proves itself intolerable.)
  • by markhahn ( 122033 ) on Monday October 28, 2013 @06:19PM (#45263395)

    systemd falls into the same trap as "desktop environments". It starts with appealing goals (basically, make startup a graph that is traversed parallel-breadthfirst), but it winds up sucking. Consider what happens when systemd dies. This happened to me recently (fedora19, upon resume) - there's not much you can do except reboot. Yes, this could have happened with sysvinit, but who among us ever had a crash of init? I certainly haven't, and I'm a certified greybeard.

    AFAIKT, the problem is that it's trying to borg a whole bunch of subsystems that do a great job by themselves. For instance, systemd tries to replace syslog for the most part. It's easy to see why it would want to do this, since daemon/server IO is a useful part of managment. But trying to do so, the system becomes more fragile and *narrower* in its applicability - more specific to how one guy (Lennart) thinks every system should behave.

    I suspect what will happen is that systemd will get shaved down a bit with some of the excess functionality removed, and in the process will become reasonably robust (ie, NEVER crash).

  • by t482 ( 193197 ) on Monday October 28, 2013 @06:57PM (#45263785) Homepage
    Size and complexity
    Upstart (1.5): 285 files, ~185k lines, ~97k C
    Debian: sysvinit + 120 files, 5.8k lines
    systemd (v44+): dbus + glib + 900 files, 224k lines, 125k C
    sysvinit: 560kB, 75 files, ~15k lines

    Debian startup is smallest, it's only shell with sysvinit (C) as dependency
    Upstart is about 10 times bigger in terms of lines of code/text. Most of the extra complexity size comes from C.
    systemd is about 10 times bigger, like upstart. But with the mandatory deps it blows up to about one hundred times the code footprint! Most of the extra code is in mandatory dependencies, but the systemd core is also bigger than anything else.

  • by phoenix_rizzen ( 256998 ) on Tuesday October 29, 2013 @05:47PM (#45273623)

    I don't care what they replace it with, so long as there's a flag/switch/option somewhere to make the boot deterministic and identical across systems and across boots.

    We run ~5000 diskless clients using Debian, booting via PXE/TFTP and mounting all filesystems via NFS.

    With Debian 5, everything ran perfectly. With Debian 7, the boot is now very racy and too many things depend on the speed of the network (some services start before their dependencies are ready). We actually had to turn on verbose boot messages in order to slow things down enough for everything to boot correctly. We're testing the "don't run in parallel" flag now to see if that fixes things.

    It's virtually impossible to debug a concurrent/parallel boot system, as every boot is just slightly different from the last. With the original sysvinit system, where things ran in series, one after another, it was very easy to find problems and fix things.

    We don't care if the computer takes an extra 15-30 seconds to boot; we boot everything in the morning via WoL before classes start, and they are rarely booted during the day. What we do care about is being able to debug problems and make things work the same, time after time after time.

    Upstart doesn't sound like it helps much in this area. Don't know about SystemD.

"Conversion, fastidious Goddess, loves blood better than brick, and feasts most subtly on the human will." -- Virginia Woolf, "Mrs. Dalloway"