Announcement

Collapse
No announcement yet.

Debian May Be Leaning Towards Systemd Over Upstart

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    With using ArchLinux I like systemd and its way of controlling services, but it lacks means to examining logs. I'm currently aware of 'systemctl status <service>' and 'journalctl'. Maybe they are enough and I'm missing the concept.

    Comment


    • #22
      Originally posted by dh04000 View Post
      So how are they going to make the non-linux ports of Debian work? Are they going to add a compatibility layer?
      Are you referring to the likes of KFreeBSD?

      Since launchd work is back commencing under FreeBSD, entitled OpenLaunchd

      https://github.com/rtyler/openlaunchd

      and seeing as their priority is to make it compatible with FreeBSD and Linux I imagine they will want to coordinate with systemd folks to see just how much systemd is inspired from OS X launchd services, and where they can make it compatible between them both.

      The more FreeBSD matures, the less I'm interested in Linux, especially seeing as with LLVM/Clang Linux has no LLDB support for ObjC 2.0, and I don't see any plans. Having that is important when one finds iOS/OS X finanically and technically enjoyable.

      Comment


      • #23
        go go go systemd!!!

        Comment


        • #24
          Originally posted by Attentäter View Post
          i don't know if any of you know this but some debian developers are also systemd developer
          Ummm... How about we instead call attention to the fact that 1 or 2 of the people on the TECHNICAL COMMITTEE-- the ones actually making this decision-- are employed by, and therefore paid by, Canonical. Much bigger conflict of interest there if you want to start down that road.

          Comment


          • #25
            If systemd becomes the official default, there is sufficient interest among developers and users that there *will* be OpenRC and/or Upstart as alternative init systems available on GNU/Linux. Most likely whatever GNU/kFreeBSD and Hurd end up using. (It's unclear yet whether GNOME would still work without systemd, and a few other bits, but almost anything else you might want on a server or desktop will be just fine).

            Comment


            • #26
              I have nothing to say against systemd. It is working fine.

              My concern is that it is getting bigger and bigger all the time and IMO this is worrysome.

              It could be a perfect replacement for sysv init without replacing syslog. I have heard that it is about to replace inetd. What is next?

              Add the news that dbus will eventually be moved to the kernel. This goes against the UNIX philosiphy that at its base is a collection of small tools specialized in doing ONE thing well.

              With systemd, Linux starts to look like WINDOWZE. So basically, what happens if systemd stops working after an update? By taking such a disproportionnate importance, I'm afraid that it could render a system unusable if something breaks.

              With PulseAudio on top of it, by centralizing everything in a single place, that sounds like a terrible catatrosphe waiting to happen. If this gets compromised, linux boxes, will become amazing spying devices knowing everything that the user is doing. I really don't like where this is going!

              Comment


              • #27
                systemd is actually really good, startup/shutdown times have been severely reduced on my end. My pc shuts down in about 1 second now (from graphical to off), it used to be a lot slower, like 5-10 seconds with normal init scripts.
                I also quite like journalctl, I somewhat do miss plain text logs but the advantages really outweigh the disadvantages in my opinion. for instance, each user can see their relevant part of the logs, so you can debug things faster (unless you are root of course, where you see everything). you can also do `journalctl -f`, which is similar to the tail command with the exception that it displays new log entries in real-time (again, good for debugging).

                Only major disadvantage is (AFAIK) that other unix like operating systems are unable to use systemd, as it uses linux specific stuff (could also be seen as an advantage, as it uses the maximum capabilities of linux).
                Last edited by AnonymousCoward; 01-17-2014, 03:35 PM.

                Comment


                • #28
                  Originally posted by dh04000 View Post
                  So how are they going to make the non-linux ports of Debian work? Are they going to add a compatibility layer?
                  That question is part of the debate. Another issue that needs to be resolved is how much the project will support non-default init systems and who within the project will be responsible for that support. However, the current proposals, if systemd is chosen as init (so far it's "tied" between systemd and Upstart, with three members in favor of each, and two more Technical Committee undecided or not yet having voiced their decision), include declaring Upstart the default on non-Linux, declaring OpenRC the default on non-Linux, staying with traditional sysv-init scripts as the default on non-Linux, and leaving non-Linux porters to figure it out for themselves.

                  Comment


                  • #29
                    Originally posted by lano1106 View Post
                    I have nothing to say against systemd. It is working fine.

                    My concern is that it is getting bigger and bigger all the time and IMO this is worrysome.

                    It could be a perfect replacement for sysv init without replacing syslog. I have heard that it is about to replace inetd. What is next?

                    Add the news that dbus will eventually be moved to the kernel. This goes against the UNIX philosiphy that at its base is a collection of small tools specialized in doing ONE thing well.

                    With systemd, Linux starts to look like WINDOWZE. So basically, what happens if systemd stops working after an update? By taking such a disproportionnate importance, I'm afraid that it could render a system unusable if something breaks.

                    With PulseAudio on top of it, by centralizing everything in a single place, that sounds like a terrible catatrosphe waiting to happen. If this gets compromised, linux boxes, will become amazing spying devices knowing everything that the user is doing. I really don't like where this is going!
                    A few details. The first one is that we are kind of assuming the UNIX way is right. I do think so, but most people just assume it because it seems intuitive. The second one, having different tools under the scope of the same project does not contradict the UNIX way, as long as those tools are modular and can work separately from the others, which happens in a good deal of systemd's modules. The third one is that if systemd (the init system) breaks, your are pretty much screwed up anyway, independently of how many other things it takes with it. Other pieces are not required by the init system to work, AFAIK, so if any other module fails, you are as able to fix it as with sysv.

                    Comment


                    • #30
                      Originally posted by lano1106 View Post
                      I have nothing to say against systemd. It is working fine.

                      My concern is that it is getting bigger and bigger all the time and IMO this is worrysome.
                      "systemd" is the umbrella name for a lot of sub projects and multiple different binaries. Its not all actually in the systemd binary.

                      Originally posted by lano1106 View Post
                      It could be a perfect replacement for sysv init without replacing syslog. I have heard that it is about to replace inetd. What is next?
                      Its always replaced inetd, that's been so for a long time. And why not? Inetd was Layer 2 handling Layer 1's (init) job. Systemd starts and tracks services, whether those services are started locally or via the net they are still services.

                      Originally posted by lano1106 View Post
                      Add the news that dbus will eventually be moved to the kernel. This goes against the UNIX philosiphy that at its base is a collection of small tools specialized in doing ONE thing well.
                      Linux doesn't have a good IPC mechanism, kdbus is an attempt to fix that. The current IPC mechanism is a hack of several different utilities that were never really meant to do what they are being made to do, and as a result they don't do the job as well as they could and don't do it efficiently. Kdbus attempts to fix that.

                      Originally posted by lano1106 View Post
                      With systemd, Linux starts to look like WINDOWZE. So basically, what happens if systemd stops working after an update? By taking such a disproportionnate importance, I'm afraid that it could render a system unusable if something breaks.
                      If your init stops working you get dropped to a (ba)sh shell with the kernel loaded, you can at least recover from there.

                      Originally posted by lano1106 View Post
                      With PulseAudio on top of it, by centralizing everything in a single place, that sounds like a terrible catatrosphe waiting to happen. If this gets compromised, linux boxes, will become amazing spying devices knowing everything that the user is doing. I really don't like where this is going!
                      How the hell did PulseAudio get pulled into this discussion? Now you're just trolling.

                      Comment

                      Working...
                      X