back to article Debian ships new 'Jessie' release with systemd AND sysvinit

The Debian project is touting new ports for ARM and POWER architectures, a bunch of software updates, an upgraded Gnome desktop and better security in its just-unleashed Jessie release. However, El Reg fully expects that the switch to systemd as the default init system will divert at least some attention from the release. …

  1. h4rm0ny

    All hail Systemd.

    And he causes all distros, the small and the great, the funded by Shuttleworth and the funded by none, and the personal distros and the enterprise-ready, to be given a package in their binary repository or in their source code, and he provides that no one will be able to boot or to start a daemon, except the one who has the package, either the name of the package or the number of his process. Here is wisdom. Let him who has understanding calculate the id of the process, for the number is that of a man called Poettering; and his process id is six hundred and sixty-six.

    1. h3

      Re: All hail Systemd.

      I cannot understand how Redhat has not noticed the amount work he causes for other people yet.

      (Think the article might have a typo when it comes to the bit about packages being the latest release because it is obviously not true as the latest XFCE is 4.12). Not being the latest release is natural for Debian.

      1. Michael Wojcik Silver badge

        Re: All hail Systemd.

        I cannot understand how Redhat has not noticed the amount work he causes for other people yet.

        Isn't that rather the point?



    So when can I run it on my Pi? Lets hope the required ARM stuff is mainstream and there is little or no optimising to be done by the Raspbian folks.

    1. Anonymous Coward
      Anonymous Coward

      Re: Raspberry

      Gnome version 3.14, should be perfect for a Raspberry Pi. As if!

  3. bailey86

    Choice during install

    Does anyone know how the choice works? Is it an easy enough choice during the install? I've read quite a bit about the choice and I want to stick to sysvinit. So, it's either stick to wheezy until Devuan is ready - or install Jessie with sysvinit.

    1. h4rm0ny

      Re: Choice during install

      I haven't tried this yet, but I understand the new Gnome has dependencies on systemd so unless you remove that, you're stuck with systemd. And I wouldn't be at all surprised if there are others in there that depend on it. So someone can correct me if I'm wrong but I believe you're basically stuck with Wheezy until Devuan is available.

    2. BlartVersenwaldIII

      Re: Choice during install

      I've had several servers running jessie since before systemd was added to the repos and I've installed a couple of new ones from netinstall. Even in the advanced install you're not able to pick your init and systemd will end up installed by default.

      Just found this page telling you how to tweak the installer to supposedly avoid this, not tried it yet myself:

      Alternatively, if you're already running systemd you can switch back to sysv via the following commands and pinning systemd-sysv at -1:

      aptitude install sysvinit-core sysvinit sysvinit-utils

      aptitude remove --purge --auto-remove systemd

      echo -e 'Package: systemd\nPin: origin ""\nPin-Priority: -1' > /etc/apt/preferences.d/pin_systemd

      echo -e 'Package: *systemd*\nPin: origin ""\nPin-Priority: -1' > /etc/apt/preferences.d/pin_systemd

      The second of those pins will block anything with systemd in the name, only the first is neccesary to stop the systemd init system being used. Fairly sure those commands should work but as ever don't attempt something like replacing inits unless you've got physical console access and a backup boot procedure, YMMV, you may experience catastrophic system failure and/or erectile dysfunction standard disclaimer warnings etc etc.

      1. bailey86

        Re: Choice during install

        Thanks for that - the instructions are also available here:

        Hmmm... what to do?

        On servers I can't really see that services such as Apache/MySQL/etc would be affected by reliance on systemd. So, it might be possible to install Jessie - carry out your instructions - and then carry on. If a required package needs systemd then use something else.

        I'm currently trying to find a comprehensive list of what packages would rely on systemd.

      2. Anonymous Coward
        Anonymous Coward

        Re: Choice during install

        A short test on an XFCE desktop clean install of jessie suggests that it turns out to be hard to remain clear of systemd on debian now. You end up constantly having to alter the pinning, things break as though you're running an early instance of testing and so on. Practically it seems a non-starter to me to run anything other than systemd. I haven't tried a server install yet, but it may well be easier. For servers, I think Devuan will get a good test when it arrives.

        Now testing Manjaro with openrc on an old laptop. What a joy. Anyone not scared about the implications of systemd's over-reach should well beyond an init system should start wondering where they want linux to go.

  4. MrWibble

    I get no results for "f*ck", I presume you mean "fuck"?

    1. Dan 55 Silver badge

      Why the article writer would be interested in how many hits uk, fuk, ffuk, fffuk, and so on get has me confused.

    2. Dazed and Confused

      Zero or more "f"s followed my "ck" surely this should return more than 616 hits?

    3. My Coat

      You know you've spent too long with linux when you read "f*ck" as "fsck"…

      1. h4rm0ny

        Um, I actually did read it as fsck at first glance. :o :(

        But then El Reg hasn't been bought out by Americans or something has it? Why the fuck would we need words censored? Self-censorship is the saddest censorship. :(

        1. Code Monkey

          The "self-censorship" (and liberal use of NSFW in other articles) is a courtesy to those employees of corporations with a puritanical view of appropriate reading.

          1. h4rm0ny

            >>"The "self-censorship" (and liberal use of NSFW in other articles) is a courtesy to those employees of corporations with a puritanical view of appropriate reading."

            Such corporations can go fuck themselves, imo. It's perfectly good and old English. It just happens to originate from before the Normans invaded and convinced everyone that if you wanted to be upper class, you had to speak French, not use Anglo-Saxon words, I've never had time for snobs, personally.

  5. Anonymous Coward
    Anonymous Coward


    Linux *never*, *never* needs a restart, it's not Windows, I've reached a 2000 years uptime, I installed my first Linux on the Antikythera computer, it was the distro Dephianos made by Ptolomeus Torvaldus (we didn't know penguins then, so we used an Ibis instead as the logo) and never, never neeeded a restart since!!!

    1. Anonymous Coward
      Anonymous Coward

      Re: "needrestart"??

      Someone has a long memory! I bet you hail from a time when windows was a pane in the glass.

      1. Anonymous Coward
        Anonymous Coward

        Re: "needrestart"??

        No, no, glass was too rare and expensive then... only those who could get the golden Apples from the Garden of Hesperides actually had glass windows, but beware of the dragon if you dared to use unaprroved apps! Windows used leather, and because of the wear and smell you needed to update it often! Also because only the barbarians - who wears boots, not sandals - used it, hence the word "re-boot".

    2. TuxIsOnFire

      Re: "needrestart"??

      I assume you don't care about kernel upgrades then

      1. wolfetone Silver badge

        Re: "needrestart"??

        "I assume you don't care about kernel upgrades then"

        +1 from me.

        Linux isn't WIndows, so doesn't need restarting. However if you update or make a change to the system it may need a restart - although not immediately. I'm looking forward to what Ubuntu will be doing in regards to updates as they've said you won't need to restart for changes to take effect. Until then though, don't update your machine if you don't want to restart. The computer will thank you for the 8 hours it spends idle without user interaction.

        1. Anonymous Coward
          Anonymous Coward

          Re: "needrestart"??

          apt-get install understand_irony

          Try, you will discover a new world...

          1. Anonymous Coward
            Anonymous Coward

            Re: "needrestart"??

            > apt-get install understand_irony

            Nah, that's now "apt-get install systemd_understand_irony" and it's a dependency so is installed by default.

      2. Dr. Mouse

        Re: "needrestart"??

        I assume you don't care about kernel upgrades then

        Actually, IIRC even then you don't need a full reboot. There is something (can't remember the name of it any more) which allows you to load a new kernel in without a full reboot, although all it really saves is BIOS and bootloader time (which can be quite a lot, especially on some servers).

        1. Anonymous Coward
          Anonymous Coward

          Re: "needrestart"??

          There are a couple of packages that can attempt that. They are not the default, ask you why. The 4.0 kernel is laying down the plumbing to achieve it - let's see when it works...

      3. Jan 0

        Re: "needrestart"??


        You must be one of those whippersnappers who's never patched a running kernel!

        Nowadays you can do it with ksplice/kpatch, we greybeards just used a debugger, for example adb* with a SunOS kernel.

        *adb - the one started by Stephen Bourne, not the TLA now recycled by Android.

    3. storner

      Re: "needrestart"??

      I understand you read ancient greek. If you try english, you'll see the article mentions that needrestart works on services, not the system as a whole.

      Perfectly normal to restart Apache, ssh and so on after upgrading libssl.

    4. I ain't Spartacus Gold badge

      Re: "needrestart"??

      My Antikythera machine ran the far superior Solaris, the operating system designed by Apollo Himself. In order to start the system you were forced to make a libation of wine, and a sacrifice of ambrosia - and since Devon's finest rice pudding was a bugger to get hold of, and no-one wants to waste good wine (or even retsina), we endeavoured never to have to do so.

    5. tom dial Silver badge

      Re: "needrestart"??

      Needrestart will notify that a new kernel is available and will only be installed by rebooting. As its main function, however, it will notify of individual services that need to be restarted in order to load new libraries or other executable parts (without reboot of the system as a whole). In many cases, though, it is as easy, and with systemd, almost as fast, to reboot the system.

      Perhaps the wish for anonymity is masking the joke icon.

  6. Dan 55 Silver badge
    Black Helicopters

    Starting from the premise that systemd is manifestly stupid and wrong (because it is, it's ballooned into some kind of intermediary kernel), how come it's on the verge of taking over practically all of the Linux distros? Is it the name of some NSA op or something?

    1. Steve Graham

      It's been designed to offer functionality to other packages, so other developers have made their software dependent on it. Young folks not brought up on the pure Unix philosophy: do one job and do it well.

      The Debian decision was basically "we don't have the resources to rip out all the dependencies in these other packages, so we just have to go with systemd".

      (Yes, I have a beard, and yes, it is grey.)

      1. John Brown (no body) Silver badge

        "The Debian decision was basically "we don't have the resources to rip out all the dependencies in these other packages, so we just have to go with systemd".

        On the other hand, considering the reach Debian and its derivatives has into the overall Linux user-base, they could have simply informed the relevant developers that systemd would not be supported so either the apps work without a systemd dependency or they don't work.

        1. h4rm0ny

          >>"On the other hand, considering the reach Debian and its derivatives has into the overall Linux user-base, they could have simply informed the relevant developers that systemd would not be supported so either the apps work without a systemd dependency or they don't work."

          That's not really how Linux distros work. Whether they include it by default or not, systemd will be available as a package to install. They can't and wouldn't stop that. I mean they could do a sort of soft block by not including it themselves, but it would be trivial for someone to create a .deb package and share it. All that would be accomplished by that approach would be mildly poorer security (have to trust an extra unofficial repo). And like I said, it's not really the Open Source way to stop people installing software if they want it.

          So systemd is available for Debian regardless. In which case, how exactly are they going to go round all those projects using it and tell them not to use it?

          Yes, I agree that Debian was the last sane person in the kingdom and now they've thrown in their lot with the lunatics because it was increasingly impossible to fight them. But I'm not sure, without major and dedicated resource, they could have.


          1. Anonymous Coward

            last sane, really?

            Then I guess Funtoo is insane on other grounds, because systemd is somewhat available but not officially supported here.

    2. h4rm0ny

      Poettering got tired of Microsoft copying UNIX, and decided to seek justice by copying them for once.

      Unfortunately, he started with their business model. :/

    3. Anonymous Coward
      Anonymous Coward


      systemd should be declared a virus, and all anti virus software should have tools to remove it. Systemd tries to infest my systems all the time - at least in FUNTOO you can set a flag that tries to strip the systemd part of applications as you install them.

  7. GrumpenKraut
    Thumb Down

    How to get rid of systemd and ban it

    1. Bronek Kozicki
      1. tom dial Silver badge

        Re: How to get rid of systemd and ban it

        Thank you for the link. I have bookmarked it.

    2. GrumpenKraut

      Re: How to get rid of systemd and ban it


      Followup: works fine; marvel at how little is output by pstree!

  8. Pookietoo

    all package versions shipping with Jessie are the latest release.

    I find that highly unlikely - surely they will be the recent version that has been tested, not necessarily the latest?

  9. pyite

    Compare this to how Ubuntu f*cked us all with 10.04 & upstart

    I upgraded my Debian system to Jessie and it is working well with systemd in sysvinit-compatible mode.

    After what happened with how Ubuntu surprised everyone by going all-Upstart in their 10.04 debacle (Upstart is still not a good idea for a production server), I was expecting the worst. However, it is doing very well so far.

  10. MadMike

    systemd a copy of Solaris SMF

    systemd is an abomination of Solaris SMF. For instance, SMF is great on large servers, where the boot time can be long. For instance, IBM P570 server takes 90 minutes to boot (no typo). SMF parallelizes the boot sequence so it is much quicker on Solaris servers. Nobody uses Linux on large 16/32 socket servers (this domain are only for large Unix servers, AIX and Solaris) which means that Linux is mostly for small servers, 1-4 sockets and on the desktop. And on such small computers, you dont need some fancy parallelizing of boot sequence. There are other advantages of Solaris SMF, that systemd does not understand and violates. Why would you build a shell into systemd?

    The point is, btrfs is a bad copy of ZFS. systemtap is a bad copy of Dtrace. systemd is a bad copy of SMF. etc etc. Heck, even Linux is a bad copy of Unix, as it is bloated and messy.

    1. billse10

      Re: systemd a copy of Solaris SMF

      have an upvote for the first paragraph. And the first three parts of the second.

      Doesn't let me say "I like this post, apart from that bit?"

    2. gibbleth

      Re: systemd a copy of Solaris SMF

      Aw, heck, I'll bite. I use Solaris every day in my job, and find it to be astonishingly bloated, fantastically slow, and very, very tedious to do simple things in.

      A fairly common occurrence in my daily work flow is copying hundreds of megabytes of logs or other data to a linux box to analyze using shell tools because a) Solaris process start time is so bad that most shell tools end up being unacceptably slow and b) Solaris default VI can't load a large file.

      Another random fact: Solaris' tar utility can't read its own output if there is a large file and you don't tell it to properly handle large files. I wish, I wish, I wish I were kidding.

      I do have much fonder memories of AIX on my RS/6000 F50, but there's no reasonable way I could actually afford a usable RS/6000 and the wife made me get rid of my 'space heater'.

      OTOH, with a bit of effort, one can afford to piece together a quad socket opteron for relatively little money, which is what I've done. It's only 16 cores and quite old, but I've got the same stack running on a 48 core machine at a place I've done work for, and working quite well.

      With modern Linux, there's very little advantage to using a legacy *NIX unless one is just deeply attached to it or to legacy software that will only run on legacy *NIX.

      1. -tim

        Re: systemd a copy of Solaris SMF

        AIX was the 1st to try this and it failed. Solaris tried this is failed. Is there a trend?

        SMF is a major reason why so many people dumped Solaris 10 (and failed to abandon Sol 9).

        However, you only need a tiny little svc.startd program to grab a contract and sleep to the end of days and the old init system is still all there (even in 11.2). Even better, in 11.2 they rewrote all the SMF scripts using a new tool which means a bit of perl script should be able to turn them back in to proper init.d scripts. A modern 11.2 system can be stripped to less than 40 processes outside of what it is supposed to be doing. With 50 processes, it can be both a parent LDOM and root zone too. I've played with system where instead of isntalling pkg://minimal-server, i used just pkg://package.pkg and it is about the smallest sol 11.2 install that I think is easy to make.

      2. MadMike

        Re: systemd a copy of Solaris SMF


        "...With modern Linux, there's very little advantage to using a legacy *NIX unless one is just deeply attached to it or to legacy software that will only run on legacy *NIX..."

        As I explained above, there are no large Linux servers suitable for business ERP systems, SAP, Database, etc. The only large Linux servers out there, are all clusters (SGI, ScaleMP, etc) and they have 100s of sockets - a true cluster. When you try to run business software, the largest servers are all Unix, and they have 16 or 32 sockets. There are no such large Linux servers aimed for business, and have never existed (google if you want). You have no other choice than use Unix or Mainframes for large business software servers.

    3. Bronek Kozicki

      Re: systemd a copy of Solaris SMF

      "Nobody uses Linux on large 16/32 socket servers"

      I think you are mistaken

      1. MadMike

        Re: systemd a copy of Solaris SMF

        @Bronek Kozicki

        Yes, you are correct in that Linux are typically run on the supercomputer top500 list, with 10.000s of cpus, or even 100.000s of cpus. I was sloppy and for that I apologize.

        I meant: "nobody use Linux on large 16/32 socket servers to run business ERP systems, such as SAP, databases, etc".

        The supercomputers are all clusters, consisting of many compute nodes on a fast switch. You simply add another node, and performance of the supercomputer increase. These servers are called scale-out servers. Scale-out servers are only fit for embarassingly parallel workloads, such as HPC number crunching, scientific computations, etc - where there is not much of communication going on between the nodes. Each node runs a tight for loop which fits in the cache, and when computation is done, the result is finally sent for aggregation and summation. Like SETI@home, which runs distributed stuff on many pc nodes. This code seldom branches, so everything can fit into the cache.

        However, I was talking about scale-up servers. One huge large server with 16 or even 32 cpus. They are not clusters, but a single large server. Business software (databases, ERP, SAP, etc) are monolithic and the code branches very heavily, going all over the place. So there are lot of communication going on between the nodes. This type of code makes scaling very difficult, and the limit is typically 16/32 sockets and this domain belongs exclusively to large Unix boxes, such as IBM E880, Oracle SPARC M6, Fujitsu M10-4s, Mainframes, etc.

        In fact, until recently, the largest Linux server I have ever seen, was a 8-socket x86 server from IBM, Oracle, Dell, etc. Now there are a 16-socket Linux out there, as far as I know. From Bull, it is called Bullion. But the scaling is awful, as Linux can not handle 16 sockets well in a scale up server. There are no SAP benchmarks this server has won. It is not even on the SAP list. On scale-out clusters, Linux scales excellent.

        Ideally to scale well, every socket need a connection to each other socket. And the number of connections increase as O(n^2). Which means that if you have 32 sockets, you will have hundreds of connections! This makes constructing large scale-up servers very difficult. Now imagine a SGI UV2000 server, which has 256 sockets. This Linux server, is exclusively used for scientific computations and no one use them to run business software. With 256 sockets, you would need 35.000 connections, clearly that is not doable. So there are lot of short cuts in the SGI UV2000 server, maybe there are only a few hundreds of connections. So SGI Linux server can not run code that branch much. Hence it is only for scientific computations.

        SGI explains themselves about their large Linux server (the Altix is the predecessor to UV2000):

        "....Typically, much of the work in HPC scientific code is done inside loops, whereas commercial applications, such as database or ERP software are far more branch intensive. This makes the memory hierarchy more important, particularly the latency to main memory. Whether Linux can scale well with a ERP workload is an open question. However, there is no doubt that with each passing month, the scalability in such environments will improve. Unfortunately, SGI has no plans to move into this [scale-up] market, at this point in time...."

        The ScaleMP has a similar server, a huge Linux server with 100s of sockets. It is also exclusively running HPC number crunching workloads.

        "...The vSMP hypervisor that glues systems together is not for every workload, but on workloads where there is a lot of message passing between server nodes – financial modeling, supercomputing, data analytics, and similar parallel workloads. Shai Fultheim, the company's founder and chief executive officer, says ScaleMP has over 300 customers now. "We focused on HPC as the low-hanging fruit..."

        No one runs SAP on these Linux clusters. SAP is for monolithic scale-up servers. On the SAP benchmark top list, there are no SGI / ScaleMP Linux scale-out servers. All the top benchmarks are Unix scale-up servers with 16/32 sockets. Fujitsu even has a 64 socket SPARC server.

        BTW, SAP Hana is a clustered database. And clustered databases have lower performance than a monolithic database, such as Oracle database. If you run a 64TB SAP Hana cluster, vs a Oracle SPARC M7 server with 32 sockets and 64TB RAM - the Oracle scale-up server will crush the cluster in terms of performance.

        Scale-up servers with 32 sockets are incredibly expensive. Scale-out servers with 100s of sockets are very cheap. For instance, one single IBM P595 server with 32 sockets used for the TPC-C record, costed $35 million. You can buy many SGI clusters for that sum. Large business servers costs very much, and the big money is there. Look at IBM incredibly profitable Mainframe business. Clusters are not profitable, it is just a bunch of PCs stringed together on a fast switch.

        >>Still don't understand why businesses buy [big SPARC M6 servers with 32 sockets] instead of scaling-out [SGI 256-socket clusters]. Cost? Complexity?

        >I'm not saying that Oracle hardware or software is the solution, but "scaling-out" is incredibly difficult in transaction processing. I worked at a mid-size tech company with what I imagine was a fairly typical workload, and we spent a ton of money on database hardware because it would have been either incredibly complicated or slow to maintain data integrity across multiple machines.

        >Generally it's just that it's really difficult to do it right. Sometime's it's impossible. It's often loads more work (which can be hard to debug). Furthermore, it's frequently not even an advantage.



        Regarding Linux being bloated. Well, Linus Torvalds himself says so.

        "...Citing an internal Intel study that tracked kernel releases, Bottomley said Linux performance had dropped about two per centage points at every release, for a cumulative drop of about 12 per cent over the last ten releases. "Is this a problem?" he asked.

        "We're getting bloated and huge. Yes, it's a problem," said Torvalds."

        1. Bronek Kozicki

          Re: systemd a copy of Solaris SMF

          @MadMike thanks for that, but I think you will agree the problem is not as much with an OS but with system (hardware) architecture and applications which run on it? Basically there are limits when you can have cache coherency between all sockets and still expect sensible performance from traditional many-synchronized-threads-paradigm applications, and beyond that you have to make a move to network (real or virtual) or message passing interfaces such as OpenMP.

          I do not quite see where is place for Linux in this discussion, to me it seems to be about discord between application(s) and hardware architecture limitations. Linux might (or might not) be a host hypervisor, guest VM or even entirely left out of picture - which would improve nothing.

    4. Adam Inistrator

      Re: systemd a copy of Solaris SMF

      "btrfs is a bad copy of ZFS" ... downvoted for this

  11. Allan George Dyer Silver badge

    So how many occurrences of 'baise' are there in the source code?

    1. Allan George Dyer Silver badge

      Sorry, messed up there, I was trying to reply to h4rm0ny.

      But, to answer my own question, 26 pages, many seemingly a typo in 'biased'.

  12. theOtherJT

    système D

    Is a piece of French slang for "système démerde" (literally "system out of shit") and means taking a cheap and easy way out instead of doing something properly.

    I find this particular coincidence amusing.

    1. Adam Inistrator

      Re: système D

      "a cheap and easy way out instead of doing something properly." sums up systemd neatly.

    2. hplasm
      Thumb Up

      Re: système D

      1000 Internets to you Sir!

  13. CFWhitman

    When you say that 'all packages shipping with Jessie are the latest releases,' I think you mean that all packages shipping with Jessie were the latest releases on Nov 5, 2014 (almost 6 months ago) when Jessie froze.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2022