On 10/22/2014 02:40 AM, Sam M. wrote:
I came across this article below, and it looks like the System D implementation on Debian has caused a great deal of havoc and alienation amongst devs and admins, so a fork is potentially in the works.
Sam, Just for the record, it's just plain old 'systemd' (no fancy caps). Perhaps a bit of background will help everyone. Just about 4 years ago Arch started the planning to move to systemd. (see Arch thread: "When will Arch switch to Systemd") About 10 months later Arch made the switch. During that period of time there was an extended discussion over the merits of systemd verses traditional SysV and the drawbacks as well. There were 2 camps, deeply entrenched, and passionate over their respective positions. As I can imagine Debian is facing, there were threatened defections, walk-outs, forks, etc... Out of the discussion, Arch saw systemd as the future (Allan is seldom wrong), but knowing there would be a few rough patches during the transition. So beginning November 2011, the transition was made. Most of the resistance and criticism usually fell into to categories: (1) the kicking and screaming that came for users that truly did not understand systemd or the benefits it could provide; and (2) those that didn't completely understand systemd, but knew enough to know the changes in administering a linux box would require a fundamental relearning of application control and logging along with the time it would take to implement that change (especially at the organization level). (Note: any time there is significant relearning required, us older dogs get really touchy about upsetting the apple-cart). And to be fair, there should probably be a 3rd category: (3) those that understood systemd, but saw no benefit to moving to it at that particular time and thought the better course would be to let it mature a bit in other distros and when benefits and stability were clear, make the move at that time. One additional thing to consider is whether the people bringing you systemd are some new, unknown group of coders touting what they think is the latest killer init, or whether the push behind it comes from the open-source leaders in the community. Freedesktop.org certainly falls into the later category (though they are not without misstep) So you throw all this into the mix and think about how it applies to Debian, a longstanding traditional distribution. There is no question that they are going through, what all distributions go though, in deciding on whether, and when, to move to systemd. If you review their discussions on it, the arguments against will probably fall into one of the three categories above. Will this lead to a fork in Deb? -- I doubt it. If the management at Debian were smarter, would they have found to transition to systemd without setting the devs hair on-fire -- of course, but nobody seems smart enough to learn from history, managers or world-leaders alike. As for openSuSE? Will there be a fork? (Ummm... WTF are you talking about?) OpenSuSE has already moved to systemd (albeit systemd 'lite' retaining consolekit for desktop user-session tracking, etc..). The only time a fork would have been possible was when openSuSE was where Debian is now. (that ship has sailed) So no, no fork here. (unless Christian is part of some underground SysV conspiracy and does something that really surprises us all). What are the lessons? If you haven't pickup up on it, then: The first would be: if you do not fully understand the benefits provided by systemd as well as the trade-offs involved -- do not whine or complain (you fall into categories (1) or (2) above.) If you fall into (3) be part of the discussion on planning and help develop a migration path that works for all devs and minimizes impact on users. If admittedly you fall into (1) or (2), you ultimately have to learn to trust the people setting the direction for the distro based on their understanding of the benefits to the distro and there ability to manage the transition. (there really are times when you just have to trust the smartest guy in the room) Finally, change does take work, but with systemd, after a steep (but short) learning curve, you find there is really not much difference from an administration standpoint. It works fine, the lingering timing issues will get fixed, there are workarounds to all of them, and in the end -- there really was no need for all the kicking and screaming that signaled the sky was falling to begin with.... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org