Although there are those who think the systemd debate has been decided in favor of systemd, the exceedingly loud protests on message boards, forums, and the posts I wrote over the past two weeks would indicate otherwise. I've seen many declarations of victory for systemd, now that Red Hat has forced it into the enterprise with the release of RHEL 7. I don't think it's that easy.
Yes, we saw systemd rise in Fedora, and we knew it was going to be part of RHEL 7. We saw systemd's inclusion in Ubuntu and Debian, and that was that -- for a certain segment of the Linux user base. The rest of us who run big Linux-based services and application stacks on CentOS and RHEL were possibly derelict in not speaking out about our disdain for systemd before these developments came to pass. But it's not too late to speak out.
[ Also from InfoWorld's Paul Venezia: Choose your side on the Linux divide | Review: RHEL 7 lands with a jolt | The beginner's guide to Docker | For the latest practical data center info and news, check out InfoWorld's Data Center newsletter. ]
I see one common thread in the heated discussions over systemd. The most vocal proponents of systemd seem to be singular desktop users, whereas server admins and architects lean the other way. An exchange on one of my recent posts showcases this. User ZincKidd comments, "BSD is looking better and better...." In reply, Adam Jorgensen notes, "Good luck running it on your laptop :-)".
With all due respect, Adam, I'm not going to run RHEL 7 on my laptop either. I'm going to run it on servers, and that is a vastly different workload. Nobody here is talking about laptops.
In another post, pro-systemd commenter Luya Tshimbalanga disparages runlevels as being "unclear." He later says the basic 0,1,2,3,4,5,6 runlevels that have been part of every Unix derivative are somehow too complex. When asked to whom those runlevels are unclear, he states: "New generation of administrators and users. You asked because you are already familiar with the init numbers term but future administrators won't without documentation. Init numbers are arbitrary and mean nothing on their own."
I don't mean to pick on these few, but they're indicative of a larger trend toward users who appear to believe that reading manuals and learning OS internals is bad, and we should plaster over all of that mumbo-jumbo with a nice, sleek -- and completely opaque -- management layer. For example: systemd.
I believe this thinking is pretty much in line with Microsoft's train of thought back in the early 1990s. This is an end-user mindset -- this has nothing to do with servers, and certainly not enterprise-level servers. This "learning is hard" mentality is very damaging for Linux as a service platform.
To pick another element out of the same comments, there's suddenly an uptick in interest in FreeBSD. I've been a FreeBSD proponent for a long time, having run FreeBSD servers for two decades now. I've heard more than a few rumblings of veteran admins exploring the possibility of migrating services over to FreeBSD instead of Linux due to systemd, and I believe this idea may find more legs as time passes. Especially now with all the fervor over Docker, if suddenly people discovered what FreeBSD jails have long been able to do, it might trigger industrial-size changes.
It could very well be that the systemd schism results in a split along use-case lines -- Linux gets the laptops, and FreeBSD gets the servers. Or perhaps we'll see the rise of a new, pure Linux server distribution that jettisons systemd and the desktop element altogether (along with the systemd dependencies found in GNOME). Of course, Gentoo and Slackware users believe they're already using that now.
The beauty of Linux and open source software in general is that this would be a perfectly acceptable course of action, if all of the parts are played right. Red Hat has likely made a misstep by forcing systemd into RHEL 7, but the outcome of this battle may be some clarity within Linux distributions. Some will favor the desktop over the server and continue using systemd, and some will favor the server over the desktop.
The downside is that there will be another plumbing framework to account for when packaging software for use on multiple distributions, but that's nothing new. The time is right for a new player to enter the game and offer an alternative to the RHEL juggernaut. Who knows? Maybe Suse can make a comeback. Stranger things have happened.
Go ahead, kids, spackle over all of that unsightly runlevel stuff. Paint over init and cron, pam and login. Put all of that into PID1 along with dbus. Make it all pretty and whisper sweet nothings about how it's all taken care of and you won't have to read a manual or learn any silly command-line stuff. Tune your distribution for desktop workloads. Go reinvent Windows.
Ultimately, that is what systemd looks like to the rest of us. It's not pretty.
This story, "You have your Windows in my Linux," was originally published at InfoWorld.com. Read more of Paul Venezia's The Deep End blog at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.
