1. 27
  1.  

  2. 16

    I know someone who worked on systemd over the summer doing some performance metrics and actually contributing some fixes. After that summer he has gone to BSD, specifically FreeBSD, citing systemd creep as the reason.

    So far he’s found bsd to be a really easy replacement. I suggest looking into trying FreeBSD especially given the large WhatsApp donation.

    1. 14

      Indeed. FreeBSD is great.

      What’s ironic is that systemd is pushing Linux towards a FreeBSD-like model. That is, a single unified core system that is well integrated and upgrades in lockstep. There are a lot of benefits to that kind of system.

      The chief issue is that Linux has not been like that ever, and it has been not doing that for a long time. From the original post:

      I do honestly believe this will end up being the start of a rocky period for Linux.

      To be clear if you’re going to stick with Linux, you will have to deal with systemd.

      In FreeBSD, you don’t have to “deal” with BSD init, or “deal” with the core system. In Linux, everything in the core is changing around, and everything that used to make sense is either gone or so different that it’s barely recognizable. It’s like operating system puberty, and like human puberty, it’s frustrating for everyone involved.

      Switch to FreeBSD, the adult operating system. Teenagers are a lot of work.

    2. 8

      The most disappointing thing about this whole mess is the complete failure of each side to engage with the other. From very start the loudest arguments on both sides were addressing strawmen or made in bad faith.

      At this point the pro-systemd camp has no reason to discuss anything with their opponents, because they’ve already won. There’s no hope of reconciliation now, and that leaves us all worse off.

      1. 8

        From my perspective (as a convert to the anti- side), there were plenty of attempts by skeptics to engage with the systemd developers. Their technical complaints were uniformly dismissed without good justifications, and nobody bothers anymore.

        For what it’s worth, countering the anti- side with some variant of “but sysvinit sucks!” is attacking a strawman. Nobody disagrees, and nobody has any particularly major complaints about the design of just the init part of systemd. People have problems with journald, networkd, its absorption of udev and ConsoleKit (as logind), etc.

      2. 8

        We need systemd tag :)

        1. 4

          So we can collectively hide it?

        2. 6

          The end is not neigh. I’m perfectly happy on Gentoo with OpenRC, eudev, Portage and the freedom to decide what I run on my own machine. There is no reason to jump ship to an environment with inferior hardware support, performance and software configurability (try enabling LuaJIT support for Vim on FreeBSD).

          1. 3

            Perhaps you meant “nigh” rather than “neigh?”

            1. 3

              I was hoping for a tie-in horse pun of some kind later on in the paragraph. :/

              1. 2

                Yes, but it’s too late to edit it now.

            2. 5

              As a user and application programmer, why should I care? I didn’t see anything specific in that article that was broken.

              1. 8

                Yup. The main people I hear complaining about systemd are Linux developers (as in, the people who make Linux). As a user, I’ve felt no impact other than complete annoyance at being exposed to this, as it seems to me, silly little war.

                I’m sure it means a lot to people. I’m sure it’s a big change and that scares people. Maybe it’s terrible. Maybe it’s not. I don’t know. I don’t care. How does it impact me? If this is really the line in the sand that kills Linux, then maybe the open source critics were right. Maybe we’ve all made a huge mistake in trusting in a system that relies on cooperation.

                Systemd won’t kill Linux. If Linux dies, it will be due to the fighting, not the technology. As the article states, it’s not that hard to switch to Unix.

                1. 4

                  This is a good example of someone who is “just” a user first experiencing systemd (and now looking at alternatives):

                  https://news.ycombinator.com/item?id=8640275

                  1. 4

                    What problems did he have with timedatectl and dbus? How would I know when I see it? How did he fix it, how would it have been easier without systemd, and how did systemd get in the way? Is his entire problem just that he’s unfamiliar with how to troubleshoot systemd?

                2. 4

                  Is there a place to get a good overview of the whole systemd debacle and what exactly everyone is so upset about?

                  1. 3

                    @ssl posted one side, here’s the other: http://uselessd.darknedgy.net/ProSystemdAntiSystemd/

                    1. 2

                      That actually looks like the most unbiased article I have seen on the topic yet.

                      1. 1

                        OK, this one is better than mine. On a side note:

                        name at least a dozen alternatives to the GNU coreutils (I can name about 7, I think)

                        What would be an example? FreeBSD, OpenBSD, Solaris, heirloom and busybox that’s what I can think of.

                      2. 2

                        Maybe this? http://lwn.net/Articles/614860/ (obviously, a bit based, though I admit I didn’t notice who it was until end of the mail.)

                        Or that: https://lwn.net/Articles/619992/

                      3. 3

                        Let me just pick on one point, please.

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

                        Well, CoreOS wouldn’t be possible w/o systemd infrastructure. Did the OP knew that?

                        1. 2

                          he did, and clarified that in a followup blog post.

                          1. 7

                            I’d be nice if it was linked somewhere. Here it is, for future reference: http://blog.lusis.org/blog/2014/11/21/a-few-things/