I don’t have a problem with different init managers. In fact, TrueOS has just deployed OpenRC as an init manager. My problem with systemd is mission creep. Statements like “sudo is a broken concept. We’re going to rewrite it for you!” and “we want to own everything between the kernel and the application layer” shows me that they don’t want to have Unix (or Linux as it stands) but they want to recreate it in their own image.
[quote=“oldgeek, post:18, topic:11067”]I’m glad you made that comment. I had always assumed that Linux was derived from Unix. But you made me look into the matter and found I was in error.
[/quote]
No, it is more of a clone. Back when Linus started writing it it was because there were only commercial Unixes (and Minix). Commercial Unixes like SunOS/Solaris, HP-UX, AIX, etc, were thousands of dollars, and the hardware to run them on (also proprietary) were in the 5-to-6 figure cost range. Linus decided to write a clone, similar to the relationship between PCDOS and MSDOS…Actually, a better analogy would probably be MSDOS and FreeDOS.
And, last I knew, it was not fully viable as an OS, more of a proof of concept. Besides, one of the original complaints about BSD was that it was a niche operating system compared to Linux. It is far less niche than openDarwin. Just sayin’.
Fair point, so far I only found systemd to give faster boot times and easier service management, I’ll worry about mission creep when it starts to cause tangible problems. It wouldn’t surprise me if the wider Linux community corrects for the kinds of mission creep which backfires - so end users don’t end up running into it.
eh, if this was directed to my comment, I was careful to say workstation/desktop BSD feels like a niche within a niche (Where desktop Linux is already fairly niche at ~2%). Of course for servers it’s a different story.
I’m sorry, I don’t mean to disparage, but I didn’t see faster boot times on Debian. I did an experiment with two Debian systems, one with sysvinit, and one with systemd. Now this was shortly after Debian started deploying with systemd.
What I have found (and this was confirmed by Jesse Smith of DistroWatch) was that the sysvinit box took 30 seconds to boot. The systemd system also took 30 seconds to boot. Jesse thought that maybe since systemd was so new in Debian, that they were still using sysvinit scripts. So he tried something that had had systemd for a while. Arch. The sysvinit system booted in 30 seconds, but the arch system running systemd actually took 40 seconds to boot.
I was just making the point that as I recall it, OpenDarwin was actually set up to not be a viable operating system, but rather a technology platform where they could reap the harvest of open source devs working on it, but not providing a usable OS. @bryanlunduke, am I misremembering this?
Strange, from boot menu to desktop with auto-login takes ~4 seconds here. I wasn’t timing startup before systemd but remember being surprised by the significant speedup. Anyone else noticed this?
One of the issues I have is that I rarely turn off my systems. My desktops, servers, none of them. So I did not see any real benefit from the conversion to systemd. So team systemd’s changes to the core functionality of how linux/unix operates are the primary visible effect of systemd, and I don’t necessarily feel that these are good things, since it is in the process of breaking compatibility with the BSDs (and anything that is not running systemd). I don’t see how, in any way, this is a good thing.
