https://lists.debian.org/debian-vote/2014/10/msg00001.html
A Debian developer submitted a general resolution on Thursday to re-open the systemd discussion. After getting the required number of seconds, the resolution entered the 2-week discussion period.
Debian's code freeze for the next stable release was scheduled for November 5th, so if the resolution passes, the freeze could be delayed.
The rational for this general resolution (GR) is: [continued after the break.]
Debian has decided (via the technical committee) to change its default init system for the next release. The technical committee decided not to decide about the question of "coupling" i.e. whether other packages in Debian may depend on a particular init system.
This GR seeks to preserve the freedom of our users now to select an init system of their choice, and the project's freedom to select a different init system in the future. It will avoid Debian becoming accidentally locked in to a particular init system (for example, because so much unrelated software has ended up depending on a particular init system that the burden of effort required to change init system becomes too great). A number of init systems exist, and it is clear that there is not yet broad consensus as to what the best init system might look like.
This GR does not make any comment on the relative merits of different init systems; the technical committee has decided upon the default init system for Linux for jessie.
Later in the resolution it does, however, note:
The TC's decision on the default init system for Linux in jessie stands undisturbed.
Related Stories
A grave bug has been introduced into the "wine" package of Debian Jessie, just days before the November 5th freeze deadline. The /usr/bin/wine launch script fails with an "error: unable to find wine executable. this shouldn't happen." message.
Debian has already suffered much unrest lately over the inclusion of systemd, with threats of a fork being issued, along with the possible cancellation of the GNU/kFreeBSD port and the possible dropping of support for the SPARC architecture. After so much strife and disruption, can Debian afford to have such a serious bug affect such a critical package so soon before such a major freeze?
(Score: 1, Interesting) by Anonymous Coward on Saturday October 18 2014, @03:04PM
I'm waiting for SN's resident systemd troll to comment on this article.
(Score: 2, Funny) by Horse With Stripes on Saturday October 18 2014, @03:27PM
Can't we just declare systemd to be "the Godzilla of the Linux community" so they can make a movie about it?
(Score: 2, Funny) by Tork on Saturday October 18 2014, @07:09PM
🏳️🌈 Proud Ally 🏳️🌈
(Score: 1, Informative) by Anonymous Coward on Saturday October 18 2014, @10:19PM
Who could have guessed that there would be a pro-systemd comment from an anti-FOSS crusader. [soylentnews.org]
-- gewg_
(Score: 2) by Tork on Sunday October 19 2014, @01:47AM
🏳️🌈 Proud Ally 🏳️🌈
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @05:27AM
Tork, just because the upmod has the same id all the time does not mean there is only one member that upmods gweg. We are ACs, we are legion. Resistance is futile. Expects us.
(Score: 2) by Tork on Sunday October 19 2014, @05:32AM
🏳️🌈 Proud Ally 🏳️🌈
(Score: 2) by cafebabe on Sunday October 19 2014, @06:07AM
I upmodded #107404 as Informative [soylentnews.org] and downmodded #107369 as Overrated [soylentnews.org] after reading the entire thread [soylentnews.org]. Even with the understanding that you was attempting to keep FOSS relevant, it still seems more like trolling in defence of proprietary software. Perhaps just don't express yourself well or perhaps I have a comprehension problem. The latter is entirely possible.
1702845791×2
(Score: 2) by Tork on Sunday October 19 2014, @06:49AM
Even with the understanding that you was attempting to keep FOSS relevant, it still seems more like trolling in defence of proprietary software.
I do appreciate you taking the time to describe your decision to moderate my comment, thank you.
I do have to admit, though, that I'm not sure why anybody would read my comments in that thread as being anti-FOSS. At no point do I disparage it, in fact I even remarked about how silly it'd sound to be against having 'free stuff'. What I did stand my ground on is that the benefit of better up-time made a lot more sense at a time when the proprietary OS required multiple daily reboots. I was consistent in pointing out that the argument hasn't aged well with time because Windows improved a great deal. I don't really see how I was negative towards OSS or positive towards proprietary software.
🏳️🌈 Proud Ally 🏳️🌈
(Score: 2) by Bot on Saturday October 18 2014, @10:29PM
He is just looking for a good excuse to link the story to systemd.
Account abandoned.
(Score: 2) by Marand on Sunday October 19 2014, @02:02AM
He is just looking for a good excuse to link the story to systemd.
I actually find those systemd troll posts in random articles amusing, because it's the classic-style kind of trolling that has mostly gone out of style as the internet has "matured". Starts out relatively logical and on-topic and then introduces insanity, off-topic, or errors and see who notices or falls for it. Much better (from a trolling perspective) than the nu-troll "UR A FAG!!!" "DIE IN A FIRE" "UMADBRO?!" style that passes as 'trolling' now. nu-trolling might seem more harsh because the messages are openly hostile, but it takes away their impact. They're weaker trolls.
In a lot of ways, the internet is a tamer place than it was; we're out of the "wild west" pioneer days and one of the side effects of that is that the trolls tend to be more domesticated. The wolves have been largely replaced by poodles.
(Score: 3, Insightful) by cafebabe on Sunday October 19 2014, @05:25AM
A good troll is subtle and indistinguishable from a genuine position. Unfortunately, what you interpret as trolling may indeed by a genuine position. The problem is that systemd starts from the Social Justice Warrior [google.com] position of "Do you hate disabled people? [youtube.com]" and one phantom dependency in one version creates a huge flap about GIMP requiring systemd. Arguably, systemd is a troll, although I fear it is a position which is genuinely held.
1702845791×2
(Score: 2) by Marand on Sunday October 19 2014, @02:29PM
My remark about "systemd trolls" wasn't about systemd itself or people advocating for its adoption. I was referring to the random AC posts that start out on-topic to the article and then segue into an anti-systemd rant. Some of them have been pretty well crafted and look like a legit discussion on the relevant article and then BAM! they're off-topic. Seriously, if you haven't seen them yet, keep an eye out. Some of them are pretty funny.
As for the authenticity vs trolling distinction, there are people on both sides that are really hard to distinguish genuine thought against trolls because they're so freaking extreme about it that you just want to assume they're troll. I'm beginning to wonder if pro-systemd people are putting up intentionally erroneous anti-systemd rants as strawmen they can dismantle, and anti-systemd people doing the same thing on the other side.
Personally, I just want it to be an optional init that doesn't have its slimy little tentacles in every other part of the system. That way I can go back to ignoring it and not having to watch every update I do to make sure some bullshit dependency change hasn't decided to try changing my init. New init systems are good, competition in any space is good. I'd be happy to let it live or die by merit, but instead, we've got Lennart cramming it into every aspect of the software stack that he touches. Like the speaker said in the video at one point, LP is involved in far too many parts of the system, and he has a tendency to link those parts instead of making useful standalone software.
(Score: 2) by Non Sequor on Sunday October 19 2014, @02:07AM
I figured that was a series of copycats rather than strictly one individual.
Write your congressman. Tell him he sucks.
(Score: 4, Interesting) by hendrikboom on Saturday October 18 2014, @03:11PM
There is hope. Maybe I'll hold off changing distros for a few weeks. It's possible that with enough time, systemd will be thoroughly debugged and prove itself. But it's entirely possible that it won't and I'll be better off without it. But it doesn't seem to be ready now.
There are packages that can be compiled, using different options, in several ways -- one that depends on systemD, and one that does not. For these packages the source code does not have a mandatory dependency, but the binary does. Presumably it would be up to the package maintainers to decide whether the dependency was essential to the package's functionality, or whether the extra functionality could be optional. They might implement this with an optional package to connect the main one to systemD. Or even two packages made from one source package, compiled with different options. All this will take time to percolate throughout Debian. And there's also the potential systemd-shim.
Gentoo, with its emphasis on user-compiled packages, makes this explicit.
Note, however, that is is possible to have systemd in a working system without it being used as the init system. In this case systemd could still continue its role in mission creep.
I have this situation on my jessie laptop, even though I use a system v init. I'm considering seeing what's left if I remove systemd altogether.
-- hendrik
P.S. I managed to write this without a single expletive :-)
(Score: 3, Interesting) by frojack on Saturday October 18 2014, @05:12PM
Agreed, no hooks into user space.
The technical committee decided not to decide about the question of "coupling" i.e. whether other packages in Debian may depend on a particular init system.
This is how a new, unproven init system should have been done, no hooks into user space, no dependencies allowed. If packages need to interface with the init system it is up to the init system to provide and maintain arms length interfaces, api calls, libraries, etc without disabling the existence of similar facilities by other init systems.
Most distros did this previously, with some distro specific wrappers around Init, some of which were heavily integrated into their distro specific tools (like Yast).
But under it all, the applications didn't need to be hacked because the distro's wrappers took care of that.
No, you are mistaken. I've always had this sig.
(Score: 4, Insightful) by Anonymous Coward on Saturday October 18 2014, @03:15PM
Systemd is a big decision with wide ranging consequences. It should not be made lightly. Many other things will need to be altered to integrate with it. As with all big changes, debate should continue as the full weight of the task becomes more apparent.
Having said all that I've been around long enough to know that systemd is already all over but the crying. These meetings are going to be little more than placations.
(Score: 2) by cykros on Monday October 20 2014, @05:59PM
And this is the problem with systemd. Ideally, a distro would come along, offer a list of sysvinit, systemd, and heck, throw in BSD init for good measure, in a nice lightbar ncurses menu at install time, and you can pick which one works best for what you're using that particular system for. I've no doubt that for SOME machines out there, systemd fixes more than it breaks (or at least, only breaks things you don't care about), but this trend of "push the software out and make it harder to remove than it is to just get used to it" is a bit unsettling. It got done in most distros with Pulseaudio (which, incidentally, I love, while at the same time, love that it was something that I installed myself, bringing me to the first time I've had it set up where it actually works right...), and for years, typing "Ubuntu" into google would bring up forum posts of "no sound" as a result of the unnecessary added complexity for a basic amount of sound output (and how many people do you think tucked tail and ran right back to Windows over that?). More complex and even integrated tools aren't something whose development I'd necessarily stand in the way of, but when they start throwing weight around and disrupting the entire Linux ecosystem as a result, it's time to be concerned.
Frankly, this all is a little too reminiscent of the Steve Jobs way of doing things. Coercion, force, manipulation, because lord knows the technocratic overlords know better what should be on your computer than you do. That may be all well and good for people just looking for a one size kind of sort of almost fits most desktop OS, but frankly, that's not why I started using Linux in the first place, and I think that's something that I'm not particularly alone in. I think the talk of forking Debian is probably pretty reasonable at this point, though it's a damn shame that it's not the systemd crowd taking the initiative to fork off a new distro rather than essentially expelling the sysvinit crowd from the ranks after having parasitically consumed the host. And this kind of hipster bullshit [forkfedora.org] as seen over on HN should sum up the whole attitude issue if anyone is still confused as to what the problem is (seriously...a bash script I can parse, but a poorly documented config file for a suite that involves binary logging and configuration? I think I'll pass).
(Score: 2, Interesting) by Squidious on Saturday October 18 2014, @03:17PM
If systemd is to be the init system of the future, let it become so by order of its own merit and not by a lack of available alternatives in the major distros.
The terrorists have won, game, set, match. They've scared the people into electing authoritarian regimes.
(Score: 1, Funny) by Horse With Stripes on Saturday October 18 2014, @03:53PM
If systemd is to be the init system of the future, let it become so by order of its own merit and not by a lack of available alternatives in the major distros.
Hogwash! Let the corporate interests be served. Corporations created Linux and have maintained it for decades, and now some individuals want to squawk about freedom of choice? If choice is so important to these individuals let them band together and create their own free OS! /sarcasm
(Score: 2) by Gaaark on Saturday October 18 2014, @05:15PM
/sarcasm orgasm!
--- Please remind me if I haven't been civil to you: I'm channeling MDC. ---Gaaark 2.0 ---
(Score: 0) by Anonymous Coward on Saturday October 18 2014, @06:24PM
Yes, you're both so spot on. RedHat isn't a corporation at all, in particular, one with the US security infrastructure as a cash cow.
(Score: 3, Interesting) by novak on Saturday October 18 2014, @03:39PM
This is all I wanted. I don't care what other people use as an init system. I just don't want to use systemd. Leaving it as an opt-out is good enough for me (though the future is still unclear).
Debian really does need to be the one to do this, because they actually have the manpower to dig systemd out of everything. This is not just swapping init systems, it is a major userland change.
I don't really use debain all that much but it's the last of the large distros to stand against systemd's takeover, and probably has more distros downstream of it than any other. This will give them the freedom to do whatever they want with init systems, even if they lack the manpower to totally rebuild from scratch without systemd.
novak
(Score: 2) by CRCulver on Saturday October 18 2014, @04:59PM
If you read the proposal, you'll see that it does not involve ripping systemd out of userland. All those desktop applications that require the systemd libs will continue to require the systemd libs. The proposal asks only that people be able to continuing running the init system of their choice as PID 0.
(Score: 3) by CRCulver on Saturday October 18 2014, @05:02PM
(Score: 1) by gcrumb on Sunday October 19 2014, @12:11AM
PID 1, rather.
No problem kid. Here - have a fencepost. I always keep one or two handy.
Crumb's Corollary: Never bring a knife to a bunfight
(Score: 2) by cafebabe on Sunday October 19 2014, @12:54AM
1702845791×2
(Score: 1, Informative) by Anonymous Coward on Saturday October 18 2014, @05:51PM
I dont know where you get debian have enough man power... It might have more man power than the average distro... but it also have much more work to do...
By no metric debian have enough power... We are under staffed in almost all areas.
(Score: 5, Insightful) by sjames on Saturday October 18 2014, @07:01PM
It's not really a manpower thing. If Debian will just refuse to expand the dependency hairball, it can be managed. If some package or another's upstream decides on systemd lock-in and an alternative package exists, make the alternative the default.
Most of the non-reactionary objections to systemd are that it creates a massive hairball that can only over time make it harder and harder to choose anything but systemd to the point that it is very likely to stand in the way of a superior solution even being created. It is exactly analogous to the way Microsoft claimed it wasn't possible to rip IE out of Windows.
The way to avoid that technical debt is to avoid taking out loans by not managing dependencies properly.
(Score: 0) by Anonymous Coward on Saturday October 18 2014, @08:30PM
This reminds me very much of how M$'s OOXML (an incomplete protocol without a single complete implementation) was shoved onto the fast-track for approval at ISO--and was rammed through.
(Fast track is a method which is -supposed- to be reserved for stuff with multiple compatible fully-implemented exemplars.)
The Debian mess: [lwn.net]
ONE GUY decided this.
-- gewg_
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @07:37PM
Thats because it was a tie-breaker vote. All tie-breakers are decided by one person. Its still a majority vote.
(Score: 0) by Anonymous Coward on Saturday October 18 2014, @05:13PM
By analogy, the Protestants and Roman Catholics seem to get along pretty well these days, even in Ireland.
That only took 500 years.
(Score: -1, Troll) by Anonymous Coward on Saturday October 18 2014, @05:28PM
people who hate systemd should just admit the real reason behind this hatred: it's not that systemd is SO badly designed or that sysvinit or any other init system is SO much superior, it's just because it's main developer happens to be a colossal asshole who thinks everyone and the universe is wrong but him. So yeah, no to systemd but at least be truthful and admit why you hate it so much.
I'm using jessie and right now the only drawback of not running systemd is not being able to run Gnome3 - this my friends is not a drawback at all. So If absence of systemd forces me to not use Gnome3, I'd say yay!! fuck them both!!
(Score: 0) by Anonymous Coward on Saturday October 18 2014, @06:07PM
My oppositions to systemd are mostly technical: I believe systemd is badly designed, and that there are init systems that are far, far superior to it. Your reasons for not using systemd may be personal, but speak for yourself only.
(Score: 2) by kaszz on Saturday October 18 2014, @06:20PM
What technical aspects of initd do you find superior?
(Score: 4, Informative) by sjames on Saturday October 18 2014, @07:20PM
I can't speak for the AC you replied to, but the big number 1 superiority is that it is easily replaced without breaking random stuff all over the system.
The old init doesn't block the development of a new and much better init. But suppose we get saddled with systemd and want to develop a nextgen init system. Well, first you'll have to not use systemd but for obscure reasons the desktop now refuses to come up. Why might that be? God only knows. Alas, udevd, ntpd, and the syslogger is also refusing to run without the hairball. Even login refuses to work. So you decide to do something in parallel. Alas, systemd refuses to even run if it's not PID 1. And it doesn't want to share cgroups with anyone else and your new init system reasonably enough wants to use cgroups.
So now you have two choices. You can either confine yourself to re-inventing systemd or developing your new init for BSD instead.
That's what people actually mean by it being monolithic. It's not that it's a single binary it's the way they are all codependent to the point that it is an all or nothing proposition.
Given that choice, I'll take nothing. SysV init works well enough for now and won't stand in the way of creating a proper solution to it's shortcomings. If they would like to do a proper design to make it truly modular, I'll look at it again and may make a different decision.
The really sad part is that at one time cgroups were supposed to facilitate extremely light weight sandboxes. But no such sandbox can be initialized by systemd because of the crazy way it wants to own everything. So if you ever want to run a system in a sandbox, you'll have to maintain sysV init, the old daemons, and login anyway.
(Score: 4, Insightful) by Hairyfeet on Saturday October 18 2014, @09:35PM
That is because systemd is the SVCHOST of Linux land [infoworld.com] and the bitch is (and why its an interesting battle) is that its NOT because somebody thinks this approach has more merit than the init system, its because red hat wants to heavily push cloud computing and Linux having its own SVCHOST that RH can stuff more and more shit into will make it easier for them.
So at the end of the day we'll see who has the real power in Linux land, the users or the corps. Please spread the word far and wide that this discussion is opened up for comments as we don't want it hidden and buried like the last one, we want to see a truly open and far debate and see who wins...either way its gonna be really interesting.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by sjames on Saturday October 18 2014, @11:07PM
I have to agree. My first thought seeing systemd was "I went to Linux to get away from craptastic monoliths".
The wedging effect reminds me of the situation with IE wedged into windows so tight even MS couldn't pull it out by the roots. And how now, years later even MS regrets that they can't make IE6 go away.
(Score: 2) by Hairyfeet on Sunday October 19 2014, @12:10AM
The reason I'm following it even though I don't have a horse in the race is this is the first time we've seen corporate interests clash with the millions of volunteers that create the actual code that the corps monetize. So what this battle will show us once and for all IMHO is whether those volunteers are being scammed by bullshit promises to work for corporate interests for free.
After all if this is rammed down Linux DESPITE the users saying in giant neon letters DO NOT WANT then we will have 100% undeniable proof that all those millions of volunteers are being played for chumps, as they have no voice, no say, they are just free labor for the corporate machine. What I do find ironic as hell is ATM its looking like Windows users actually have more say in the direction of their OS than Linux users have! The PTBs tried to force Metro and what happened? The users exercised the power of voting with their wallet and now Windows 10 is what everybody wanted, Windows 7 with a few tweaks. Linbux users have likewise said DO NOT WANT but without the power of the wallet? Looks like RH is just gonna ignore the users and keep on ramming home systemd.
In any case its gonna be interesting to see who wins, worth breaking out the popcorn and watching the fight.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by kaszz on Sunday October 19 2014, @01:54AM
If you can gather enough developers and support in the form of compatibility and end user hand holding. Then you can also pull the rug under Red Hat or any other corporate troll.
(Score: 3, Insightful) by Hairyfeet on Sunday October 19 2014, @02:58AM
Sorry but that is a myth with NO basis in reality, its like saying "Well if you don't like the way NASA is going you can take this iron ore and make a Saturn V, after all the plans are FOSS" which while TECHNICALLY true ignores the simple fact that what you are suggesting can only be accomplished by the rich who can afford to buy enough developers to pull that off, see mark Shuttleworth for a perfect example of that.
But to tell the end users "just fork it" is as delusional as me handing you the plans to the MIG 15 BIS (available online) and telling you to have me one built by Thursday. source code isn't magic, it requires hundreds of VERY skilled and NOT at all cheap developers to fork anything of that size and complexity. Hell if RH gets their way it would easily take dozens of highly skilled low level OS coders to rip out the systemd infection, it would be about as possible as me telling you "Hey here is the source code for Win2K, rip out the SVCHOST code and add 64 bit support while you are at it, I'd like it by the end of the year"...is it POSSIBLE? Sure but so is you winning the powerball 5 times in a row, will it happen? NOT A CHANCE IN HELL.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by kaszz on Sunday October 19 2014, @03:09AM
So there isn't hundreds very skilled developers outside of the reach from the Red hat empire?
(Score: 2) by cafebabe on Sunday October 19 2014, @06:12AM
There are plenty of developers. The problem is to get a large number of them to work harmoniously on one project. Paying their bills works for a fair proportion. However, that reduces the problem to having a large reservoir of funds and/or a large cashflow.
1702845791×2
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @08:00AM
Actually the resources you speak of are needed for maintenance. The initial fork, esp with debian, is probably tractable enough without that level of resources. Then it boils down to a popularity/utility race to see who has enough resources to maintain what kind of a distro.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @07:59PM
If systemd is as bad as claimed, then there will be no shortage of skilled developers volunteering to maintain a fork, because its what they'll be using, and many will be grateful such a thing exists. If, however, there's nobody volunteering to work on an alternative, that proves that systemd is either the superior choice or its equivalent to the alternatives, or at least not any worse.
(Score: 1) by Squidious on Monday October 20 2014, @01:08PM
Your last statement does not hold true in the following case: Nobody is working on an alternative to systemd because sysvinit is already sufficient to our needs.
The terrorists have won, game, set, match. They've scared the people into electing authoritarian regimes.
(Score: 2) by Hairyfeet on Monday October 20 2014, @03:49PM
But if RH has their way (and it is looking like they will) then there will be a hell of a lot more than init tied into systemd which means you be fucked. From the looks of it they really do want something as large or larger than Windows SVCHOST, they want systemd to be the bootstrap AND a hypervisior AND and network bridge AND lord only knows what else.
So what good is sysvinit if you still have to have systemd to have a functional OS?
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 1) by Squidious on Monday October 20 2014, @07:54PM
I was replying to the AC above me, I should have quoted. The nesting gets lots after it gets too (10 levels?) deep.
The terrorists have won, game, set, match. They've scared the people into electing authoritarian regimes.
(Score: 2) by Hairyfeet on Monday October 20 2014, @09:15PM
They really need to adopt a system similar to Ars Technica where the relevant post is quoted because the current design? Seriously fucked up. I have had posters accuse me of answering myself when in reality I'm answering three different questions put to me by three different posters in the same thread just because the fucked up formatting put all my responses at the end, its completely whack and makes anything more than 2 posts deep damned hard to follow.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by sjames on Sunday October 19 2014, @08:15AM
You'll need to watch for a good while to see how it goes. Unlike proprietary systems there remains the possibility that someone will get fed up enough to create an archive of higher priority Debian packages that have systemd ripped out or even sabotaged. Even in Debian, it is not thus far nearly as mandatory as the systemd people would like. The current resolution proposes to make staying that way a matter of policy.
I would consider ripping systemd out of Gnome myself except that a series of terrible design choices by the Gnome team convinced me to move to xfce4 already (note, switching out the GUI wasn't really an option for Windows users, the best they could do is add some hacks to it).
It's not easy but it is possible for a significant group of users to abandon the major distros and make Linux what they want it to be.
(Score: 2) by Hairyfeet on Sunday October 19 2014, @09:21AM
I said it was possible, just as its possible to win the powerball multiple times, but just as no sane person would tell someone to magically fix their debt by "investing" in powerball so too is just fork it a lie for a project of any complexity because the odds of successfully accomplishing both are roughly equal. Hell look at Openindiana, that is made up of guys that actually built Solaris and even THEY haven't been able to get to a stable 1.0 release. For an example of what you end up with see ReactOS, which in case some don't know was originally call FreeWin95 which just illustrates what lack of finances does to a release schedule.
So I'm sorry but its just not gonna happen, you just can't come up with enough highly skilled coders with enough free time to rip out something that complicated, especially if RH has their way and makes systemd into SVCHOST for Linux.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by sjames on Sunday October 19 2014, @09:54AM
So Slackware and Debian are myths? For that matter, at one time RedHat was a tiny operation.
Or the many many distros derived from them?
Gentoo is an optical illusion?
I'm not claiming it's not a major undertaking, but it has much better odds than powerball.
(Score: 2) by Hairyfeet on Wednesday October 22 2014, @01:48AM
Slackware,Debian, and Gentoo were all created when Linux could fit on a single floppy and thus got to have the advantage of years where 3 or 4 guys could maintain the entire thing. this let the community grow organically over time...that is not possible now.
If you want to see why just look at the gaming industry, in the Wolfenstein days you have all these shooters made by a couple guys in their basement...why doesn't that happen today? Because back then the entire game could fit on a floppy with a couple other games, that's why. When you can bang out the AI in a weekend you have a huge advantage over those that come later and require 10k LOC to get their AI up and running because again you require much less resources. And before you link to somebody's homepage I'm talking about games that can compete against the big guys NOT some guys project that looks like it was made on the SNES.
There is a reason why the only major player to arrive on the scene since those early days was backed by a millionaire and that is because the complexity of building your own distro from scratch, i mean ACTUALLY building your own distro, not just taking others packages and sticking them together with minimal alteration? is EXTREMELY complex and requires a shitload of guys that the odds are you just won't get to work for free. Is it possible? Sure but so is powerball, hell I could wake up in the Matrix tomorrow, anything is possible...but is it likely? Not in the slightest, sorry.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by sjames on Wednesday October 22 2014, @03:46AM
For one thing, linux didn't exactly fit on a floppy. You at least needed a boot floppy and a root floppy. But when SLS morphed into Slackware, it was 30 or so floppies.
You seem to be missing a BUNCH of distros. I presume you're talking about Ubuntu as the latecomer. But you forgot Gentoo, Arch, Mint, Scientific Linux, Rocks, Kali, etc. etc. etc.
The part you forgot is that a new distro gets to start with an existing one. There's no need to start from scratch (though Gentoo did).
I know what it takes to maintain a distro fork, I have actually done it.
(Score: 2) by Pav on Sunday October 19 2014, @01:51AM
My personal objection to default systemd would probably go away if the Unix world (not just Linux) decided that the init standard needed updating, and the hard yards were put in to implement a robust standard. This way each piece of systemd would have to implement a well defined part of the standard, as would any replacement of that piece. A defacto standard should NOT be acceptible for something as central as init, especially if implemented as a monolithic hairball. It relegates any competitor to be a perpetual second-mover (see Samba vs Windows before the standards).
(Score: 2) by opinionated_science on Saturday October 18 2014, @07:48PM
I'm kinda agnostic, though having seen LP talk, he is clearly not a dumb fellow.
The only thing about systemd I don't like is the fact is it is "opaque" in a way that init , really isn't.
I also understand WHY systemd is being pushed, because coupled with BTRFS and volume management, linux is poised to become considerably more productive an ecosystem.
The thing is evolution is sort of slow, until we know the right way ahead. But there is definitely a bigger plan than just systemd, and LP is happy to tell anyone who will listen about it.
I stopped complaining when I saw how *many* projects he is involved in.
Don't shoot the messenger, even if he is a bit a bit (abrasive/aloof/impolite/arrogant)*
*delete as applicable
(Score: 0) by Anonymous Coward on Saturday October 18 2014, @08:28PM
>linux is poised to become considerably more productive an ecosystem.
>there is definitely a bigger plan than just systemd, and LP is happy to tell anyone who will listen about it.
I for one do not care if RedHat wants to become a new Microsoft and bless us with a new Windows RT (called Linux RH this time).
But I very much do care about being unable to opt out of being so blessed.
They are welcome to have their "ecosystem" to themselves, and play with it as they wish. They are NOT welcome to force OUR ecosystem into a clone of theirs.
If their solution really has technical merit, people would adopt it by free choice. And what the dirty politicking really means, is that devs themselves do not believe in that, and neither do their bosses.
(Score: 2) by Geotti on Saturday October 18 2014, @10:58PM
I stopped complaining when I saw how *many* projects he is involved in.
If you'd have a closer look, you'd see that only once he steps down from projects do they actually get the chance to become somewhat usable.
(Score: 2) by sjames on Saturday October 18 2014, @10:59PM
I like btrfs in concept, but it simply isn't mature enough yet. It has improved a lot, but it's response to a failed drive in a redundant setup is data threatening.
I feel confident that will be worked out, but not today.
I really don't find the larger plan all that compelling at this point. It struck me as an elaborate and resource hogging way to avoid playing nice with others. The thing is, with an appropriate design systemd is entirely unnecessary for that to work (if you want it to work).
I would go so far as to say that none of the things about systemd that worry people are at all necessary to achieve the stated end goal AT ALL. The only thing they are necessary for is wedging systemd in too deep to make digging it out later practical. The designers are either idiots or crafty in a bad way.
(Score: 1) by linuxrocks123 on Sunday October 19 2014, @12:27AM
He's been involved in several projects, and they all suck, usually even after he leaves. They also (until he leaves at least) have an actively hostile attitude towards every OS that is not Linux. He ignores standards, disdains lifting even a finger for cross-platform compatibility, and has no respect for the importance of modularity. He is an arrogant, immature twat with quite the superiority complex. There's a recent video on YouTube where he continuously interrupts and argues with a speaker at a conference, culminating in him actually getting up on the stage uninvited to answer questions. Until he grows up, if he ever does, he needs to be steered to work on unimportant shit like phone apps where his provincial and short-sighted attitudes can't do much damage.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @03:05AM
Is this the video where he accuses the guy who disagrees with him of hating the disabled?
(Score: 1) by linuxrocks123 on Sunday October 19 2014, @05:13AM
lol yes.
If you have a link to it, please let me know.
(Score: 2) by kaszz on Sunday October 19 2014, @01:59AM
In that case we still need something other than systemd, but not initd either. So a third way is needed to improve the ecosystem.
(Score: 2) by forsythe on Saturday October 18 2014, @08:06PM
As the AC you're replying to, I'll just point to Rich [ewontfix.com] Felker [ewontfix.com]'s analysis.
(Score: 2) by forsythe on Saturday October 18 2014, @08:20PM
And I submitted instead of previewing. Some points I wanted to include in that post:
1) I didn't post as AC because I included links that I thought may be interesting.
2) I find any other init system superior by virtue of not including the flaws that systemd includes, particularly
2.1) The flaw of future lock in. On a machine with systemd, transitioning to any other init system is a pain in the neck. Sysvinit, Epoch, etc. are replaceable with far less effort. When PerfectInit 1.0.0 is released, I can transition to it without recompiling my whole system - at worst I'll have to transition a few scripts. Systemd is not PerfectInit, and from its design choices I doubt it will ever be PerfectInit for me. Therefore choosing systemd now assures me further pain in the future. Sysvinit is therefore superior to systemd for my purposes
2.2) The flaw of complexity. As far as I'm aware, any other init system is less complex than systemd. Since complexity correlates strongly with bugs (and systemd is not one of those special projects that have been formally verified), any other init system is more likely bug-free than systemd, and therefore superior for my purposes.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @12:57AM
as a debian user, my init of choice is sysv. the only systemd dependency that i can find seems to be libsystemd-login0, which is a dependency of dbus, but it's the dependency on dbus by so many other programs that makes it difficult (not impossible) to avoid systemd altogether (dbus is the only package that i have installed that depends on libsystemd-login0).
as a test, i was able to run a system without dbus by installing openbox and using gentoo package (not the distro) as a file browser. many programs broke, but i was still able to use hexchat and other utilities like synaptic.
in the end i choose to keep the libsystemd-login0 package installed along with dbus because i like using programs that depend on dbus (like xfce and nautilus).
pid 1 isn't systemd, and if the packages that libsystemd-login0 replaced weren't being maintained then it seems like a natural progression. perhaps if the libsystemd-login0 package maintainer had simply named the package 'lib-login0' nobody would even notice or care.
i don't feel any more pressure to use components of systemd than i do something like say xorg
at the end of the day, if it works i'll use it
i also run debian as a vbox on a windows 7 host... does that make me evil?
(Score: 2, Insightful) by kaszz on Saturday October 18 2014, @06:17PM
System design decisions should be taken on merits of the design itself not the personality of the designer. The designer can be replaced while still using the system. But replacing the design gives you another system.
Systemd seems to have too much dependencies ie single point of failure, expose more attack surface than needed, destroy userland API compatibility, complicate investigation to the reasons of system problems etc. The design is not good enough.
This puts the grounds for the project management to adopt systemd into question. Is it on strict technical merits or is there other less honorable reasons?
(Score: 2) by Pav on Sunday October 19 2014, @02:15AM
If systemd really fixes problems worth solving there needs to be a standard agreed upon across the Unix world. Without this systemd has no place as default init, or anything else really... init and intrasystem communication is no place for a defacto "standard" (see Samba vs Windows networking before the SMB/CIFS standard).
(Score: 3, Informative) by rufty on Saturday October 18 2014, @07:34PM
I'd never heard of him until this, and some of what systemd does sounds both a good idea and needed. But binary logs? NO. JUST NO.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @12:32AM
I don't see what's the big deal with binary logs. The log file itself already is stored on a binary filesystem. Do you object to binary filesystems too? If not, then what's so inherently different in formatting a bunch of data records in a log file, as opposed to formatting a bunch of data records in a filesystem?
(Score: 0) by aos on Sunday October 19 2014, @02:12AM
Text is more resilient without sophisticated recovery tools in the event of corruption.
I've used binary logs on embedded systems and they were great for preserving log history and minimizing the impact of collecting logs in real time while testing (less data over the bus!). However if there was a mismatch between the loaded build and the translation file to get human readable text, you would get garbage which could be a pain. Also, if you start getting data in the middle of a log line sequence, or one of the bytes got corrupted, you would lose those entire entries (possibly many depending on what sequence you lost). Sometimes they contain critical information.
With respect to the file systems, even if you get corruption on the file system metadata, you can still in theory pull the raw data down which should be stored as plain text in ext4 and find it (I think?). If the file itself got corrupted, it probably doesn't affect every byte, so you can still make sense of it easily with minimal loss of information (possibly none of it only affected static string portions). Maybe it would be different on something like zfs.
(Score: 2) by cafebabe on Sunday October 19 2014, @06:30AM
\n is a perfectly good record separator and it allows graceful recovery. A valid exception is packet logs but that's only because the original data is already in a binary format. Even then, escaping should be used so that a magic sequence can be used to allow graceful recovery.
Printable logs simplify legal aspects such as best evidence [wikipedia.org] and chain of evidence [wikipedia.org]. If someone gained illicit access to a computer, I wouldn't want to argue about hexdumps in court or see a case dismissed due to discrepancy, technicality or needless mis-understanding.
1702845791×2
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @08:22AM
"graceful recovery"? Please specify/clarify.
"Printable logs"? There is plenty of code between the log bits and the printer. So what if some of that code turns a differently binary coded logfile into the same sort of human readable form? Perhaps better suited to the written language of the jurisdiction where it happens?
Thanks for the 'best evidence' link. Got me some education. But I don't buy it. It sounds like an appeasement of the clueless sort of judge who considers a text font foreground and background color as a means to bypass 'equal justice under the law'. I'm kindof hoping that post-snowden that shit don't fly anymore. Likewise whatever technically clueless judges your sentiment is meant to appease, ought to be laughed out of the system after a supreme court ruling with half a clue gets the legal system into the 21st century.
Seriously though, english-centrism, and avoiding it, is what persuades me that in many cases a binary log format is worth the inconveniences. That said I'll probably be clinging to non-binary format logs in many cases because english is my first language, and there are conveniences I've grown accustomed to.
(Score: 1) by dlb on Sunday October 19 2014, @04:15PM
But binary logs? NO. JUST NO.
Granted, binary logs isn't the main point of contention about systemd, but it's consistently brought up in these type of discussions. So I'm curious, would it be hard to tweak it to output textfile logs? Doing so might win a convert or two to the pro systemd crowd...not that I'm advocating one way or the other.
Anyone know about the possibility?
(Score: 1) by ansak on Sunday October 26 2014, @05:22AM
Those who support systemd are quick to exclaim (truly) that there is: here's arch's way [archlinux.org] of doing it, for instance. It is disabled by default but it's there.
cheers...ank
still looking for Gentle Treatment
(Score: 1, Interesting) by Anonymous Coward on Saturday October 18 2014, @05:40PM
Where in the Debian by-laws does it say the TC had the authority to make this decision to begin with?
(Score: 2) by Geotti on Saturday October 18 2014, @11:15PM
According to Bdale Garbee [wikipedia.org]
We exercise our power to decide in cases of overlapping jurisdiction
(6.1.2) [...]
[https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=727708#6236]
From the Debian Constitution [debian.org]
The Technical Committee may:[...]
2. Decide any technical matter where Developers' jurisdictions overlap.
In cases where Developers need to implement compatible technical policies or stances (for example, if they disagree about the priorities of conflicting packages, or about ownership of a command name, or about which package is responsible for a bug that both maintainers agree is a bug, or about who should be the maintainer for a package) the technical committee may decide the matter.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @02:26AM
Ie nowhere. The tc is for deciding bug fix when two package maintainers can't agree. Not for changing fundamentals of the os. It was never intended for this. It was only intended to fix simple straightforward issues between two simple packages.
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @07:40PM
You are free to work with the other developers to create and maintain the alternative.
(Score: 2) by Geezer on Saturday October 18 2014, @10:42PM
Much discussion lately in the Slack world about the supposed inevitability of an init change in Slackware, with polls suggesting that there is no good reason (as yet) to do anything but keep BSD sysvinit. If Debian wants to wander off into systemdland, fine. At least we have BSD and Slackware to keep the Unix stuff, um, Unix.
Vaya con "Bob".
(Score: 2) by yellowantphil on Saturday October 18 2014, @11:43PM
I haven't kept up with Slackware lately, but last time I checked, not even Pat [linuxquestions.org] would rule out switching to systemd. It seems like the strongest holdout left is Gentoo, but I don't want to use that distribution. (I run Arch, but I don't say anything on the Arch forums about systemd, for fear of getting burned at the stake for my lack of enthusiasm.)
I don't understand why old UNIX people like Pat and Linus seem fairly neutral to systemd. Running jobs in parallel while booting is cool, but you can even do that with bash scripts if you're a bit clever. What about systemd trying to take over everything in the OS, to the point that desktop environments can't even run without it? Why are so many people not concered about this?
And... how did Soylent News become a hotbed of systemd hate? I'm not complaining, but it feels like everyone who dislikes systemd came here.
(Score: 1) by novak on Sunday October 19 2014, @12:33AM
I might recommend CRUX as a decent non-systemd distro. Apparently, it was a lot of the inspiration for the much more popular Arch linux. You do have to compile from source rather than getting binary packages but everything is dead simple. I've been using it since 2010 as my distro of choice- it's probably the most unix-like linux I've ever used, including slackware.
As to why this is a hotbed of systemd hatred... That's a good question, I've noticed it too. I didn't realize that it was until after joining, but I was pleasantly surprised. My guess is that a lot of people wanted to leave *cough* thatoneothersite *cough* as it was being ruined by corporate interests, and thus are likely to feel that what systemd is doing to linux is also a hostile corporate takeover.
novak
(Score: 2) by cykros on Monday October 20 2014, @09:31PM
The comparison between the course taken by distros with systemd to what that other site did that brought us all here in the first place is a very apt comparison (no Debian pun intended). It's a matter of trying to lure in the masses by alienating their already established userbase (many of whom are using Linux in the first place because it's thus far generally been very much NOT one-size-fits-all, but takes a more modular approach that lets you mix and match components). Poettering's statement that "Linux is still too fragmented...[and] needs to be streamlined..." is more than a little worrying for those of us watching more and more get tied into his creations. While it's thus far been packages I really can't say I miss (Skype, Gnome...), I can't say I'm looking forward to the day that it's something a little more mission critical getting caught up in the mess.
It's one thing to craft software into an ecosystem to bring reality to your vision. It's another to go around in the software equivalent of ethnic cleansing in order to do it. Android went and did a whole lot differently than other distributions of Linux, and has been a generally welcome enough addition to the family (even if it isn't everyone's favorite family member, I think we can all applaud the additional hardware support and other contributions it has made back to Linux in general). Ubuntu Touch, and to a lesser extent, Ubuntu, have been making all sorts of their own little systems in order to bring about their vision, and somehow, have managed to generally do so without demanding that the rest of the community use their way or face irrelevancy. And again, not everyone's favorite family member (definitely not mine), but we can generally appreciate the additional interest in driver support and things like Steam having been finally ported over so the rest of us can use it. Systemd isn't like that. It comes in, takes the software you've been using for ages, and says "if you want to keep using this, you'll have to stop using whatever init system you were, and start using ours...you were doing it wrong in the first place anyway". I don't take kindly to that kind of attitude from humans in most any arena, and I sure as hell don't appreciate it coming from software. Linux being a place where I'm free to do what Poettering has in mind isn't the part of Free Software that brought me in in the first place, and I'd question anyone okay with the attitude, regardless of how much you like or dislike systemd as a piece of software otherwise.
In the mean time, I applaud the efforts of those ensuring that there are still alternatives being developed, packaged, and used throughout this ordeal. As a community, providing alternatives has been one of the things the Linux world does best, and provided that remains the case, this should be something we can get past without too much of a disaster. Bonus points to out to whoever finds an elegant way to backport systemd dependent software to other init systems!
(Score: 1) by Bill, Shooter Of Bul on Sunday October 19 2014, @12:46AM
I think old unix guys don't hate systemd, because they are a lot wiser than everyone else. They understand as even the original unix developers understood, that it was a good design. They did not believe that it could not be improved upon. So they are evaluating it on its own merits, without holding unix up as religious doctrine.
(Score: 2) by dmc on Sunday October 19 2014, @01:54AM
mod parent comment up. Though feel free to mod mine down as I'll add that beyond completely agreeing with the comment, I *also* believe this is big league power politics. I think the old unix guys are so familiar with the days when you had to be a TLA to afford a unix box that they don't feel the need to dwell in the crossfire of that psyop battle raging with god knows how many falseflags flying around in the flamefests masquerading (well I might add) as technical debate. (disclaimer: high thc levels in blood :)
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @03:01AM
You forgot to add a disclaimer mentioning that this is only your opinion and nothing more.
(Score: 2, Insightful) by GeminiDomino on Sunday October 19 2014, @04:06AM
Claiming that the opposition to systemd is based on some sort of cultish, slavish devotion to "unix principles" is nothing more than a disingenuous handwave. Just the handling of the logging alone is enough of a technical failing, and that's without taking into consideration that it's an all but incontrovertible sign that the designers don't actually understand the use cases of the systems they're developing for.
"We've been attacked by the intelligent, educated segment of our culture"
(Score: 2) by dmc on Sunday October 19 2014, @07:12AM
I didn't read BSOB's comment in that light. BSOB didn't seem to suggest that was the "basis" of opposition to systemd, but rather just something that has come up in the discussions. Which it has, though s/cultish slavish devotion/clinging attraction/ (for some of the instances). I think you need to focus less on those words, and more on BSOD's general point of - as I would paraphrase - "its some code, if its useful to some folks, it'll get used".
Likewise I'll highlight what I think the core of your problematic opinion is in my opinion. You say-
But that sentence doesn't make logical sense to me. I believe the designers of systemd are developing for a set of use cases that they themselves have arbitrarily chosen. They are designing software exclusively for their own purpose. It's not about you, it's about them. The wise will sit back and hope that the "some code" they write happens to at some point become useful to them, and be used for their purposes.
Mind you, I think from what I've read, were I a heavier debian user, I would vote with Ian Jackson, if for no other reason than to irritate those that his view seems to be irritating. But at the same time, if 5 years from now, I'm putting together a system and some of this systemd code seems to help in some way or another, I'll be glad of it. I think it's likely to happen. Despite the fact that I haven't yet found a compelling need to run it. But as it matures, I keep an eye on it, figuring it may help me out somewhere somehow sometime. It can't hurt. If debian goes systemd only, I know there will be other choices elsewhere.
(Score: 1) by GeminiDomino on Monday October 20 2014, @01:47PM
I believe the designers of systemd are developing for a set of use cases that they themselves have arbitrarily chosen. They are designing software exclusively for their own purpose. It's not about you, it's about them. The wise will sit back and hope that the "some code" they write happens to at some point become useful to them, and be used for their purposes.
No, there's no use case where the way they handle logging (not just the stupid binary aspect, but the "by-design" loss-to-corruption) is anything but incompetent.
And you have a strange definition of "wise." Not everyone's sole exposure to Linux is a toy hobbybox that they can afford to have blow up on them while they wait for "some code" to become useful to them. The REALLY wise ones need to find alternatives before the garbage ends up infecting any actual production machines.
"We've been attacked by the intelligent, educated segment of our culture"
(Score: 0) by Anonymous Coward on Sunday October 19 2014, @04:09PM
Yeah, we understand that SysV init sucks, and we've known about that for 30 years.
The whole anti-systemd thing is like #GamerGate but instead of hating women, they hate unix. But they promise they love it.
Mostly they were told by their "friends" that it is this devil thing they have to stop, so they hate hate hate. Nevermind that their reasons for hate are a mixed bag of lies and misunderstandings...
(Score: 2) by cykros on Monday October 20 2014, @09:35PM
Old Unix guys don't hate systemd because it only runs on Linux. Half of them haven't even seen it come up as a blip on their radar, as the BSD's all still have their own systems in place and are pretty out of harm's way. When software that was cross platform between BSD and Linux starts coming up as only working for Linux with systemd though, some might get a little annoyed...
Though the other big reason: old Unix guys spend less time complaining and more time coding. Poettering seems almost to flourish in the dissent; it's people coding things in a way he doesn't envision that gets him bothered. Bit of a Steve Jobs complex in that regard.
(Score: 2) by cafebabe on Sunday October 19 2014, @07:10AM
If you read Neal Stephenson's comparison of Unix to a commercial power drill [team.net], you'll understand that some of the Unix greybeards are extremely pro-choice. If you want to type sudo rm -rf / then that's your prerogative. Likewise for stripping out init.
In some cases, you only have to suffix an ampersand. However, I'd prefer my systems to at least attempt to boot deterministically and in a stable manner.
People got banned from discussing it in appropriate venues so they took it elsewhere. Quenching discussion seemed to be effective until it was revealed that a mutual dependence with Gnome3 was driven by [youtube.com] political correctness rather than technical merit [google.com]. At that point, it shared the moral lowground with #gamergate.
1702845791×2
(Score: 2) by cykros on Monday October 20 2014, @09:47PM
I wouldn't say Pat seems all that neutral; he just has made sure to be clear that he's not zealously opposing systemd, but rather, expressing a hesitancy towards it due to not seeing any good reason to incorporate it into Slackware at this time (especially considering that Slackware has by far the prettiest implementation of SysVinit I've come across anyway, what with using BSD style init scripts with it instead of the mess that other distros generally had going on before the switch). Frankly, not surprising to see from him, as he's shown himself to be a pretty bright beacon of sanity now for over 20 years.
It's all about weighing the pros and cons, and I trust that it'd take quite quite a good set of reasons to sway him into throwing it onto Slackware. Especially at this point, given how much more interest there has been in the use of Slackware over the last year or so generated by the refugee influx. Though I will say that as a Slackware user, I wouldn't be terribly upset to see systemd pop up on Slackbuilds.org or in one of the other third party repos as an option, if for no other reason that I'm not so much anti-systemd as I am pro-choice. As soon as embracing systemd without wavering on that stance becomes a little less oxymoronic, I'm sure y'all can expect me to pipe down already :-).
(Score: 2) by cykros on Monday October 20 2014, @10:01PM
I'm finding it ironic that it actually was Gnome3 that got so much of the Linux world in a huff, because the only ones who have liked anything about Gnome3 seem to be its own developers anyway.
All that said, it is a worthwhile enough canary to pay attention to. I don't think anyone wants to see a day when something like OpenSSH or Apache suddenly has systemd as a dependency, so while no fucks about Gnome3 specifically are merited, it's perhaps best to remember what happened when we took too long to react to Hitler's actions in the Rhineland. And no, I'm not saying Poettering is Hitler, even if he is a German that thinks things should be done his way or else and has a knack for manipulative rhetoric...just that it rarely pays to ignore hostile activity, even when it's towards a member of your community less liked than others, particularly when the source of that hostility has already expressed a desire to perform similar acts against other community members, and generally obstruct self-determinism wherever possible in favor of fostering monoculture.
(Score: 2) by cafebabe on Monday October 20 2014, @11:40PM
Ein volk, ein Führer, PID ein!
(One people, one leader, PID 1!)
1702845791×2
(Score: 1) by quixote on Sunday October 19 2014, @01:59PM
and even I'm against it from what I've been hearing. I'm on Debian testing, which I'm pretty sure uses systemd now. The only obvious difference to me is my laptop boots faster, which is nice, but ...
1) Binary logs? No. Even I've used text logs for troubleshooting. They're incredibly useful and wonderfully transparent. Why would anyone want to move away from that? Except if less transparency was actually the point for them.
2) Total dependency of everything on this one process to run? No. Single points of failure are always stupid. Unless somebody is looking for more control.
3) I seem to remember that the same developer who did pulseaudio is big in systemd? Hello? I was using ubuntu when pulseaudio became a big deal and I was a regular on their forums.Every second question was about a pulseaudio problem. (I exaggerate, but really, there were a LOT of issues.) That's not lethal for audio. But it's not a good idea for PID 1.
4) The big push behind this is RedHat wants to cloudify everything? God save us. Nothing wrong with cloudiness in principle. I even have a little internet accessible server. Very convenient. But when a corp wants to put you on a cloud it's always called Hotel California. Do. Not.Go.There. Run for the hills! You hear me, Debian?
(Score: 2) by cykros on Monday October 20 2014, @10:15PM
2) Total dependency of everything on this one process to run? No. Single points of failure are always stupid. Unless somebody is looking for more control.
While I'm generally not in the systemd apologist camp, it's worth pointing out that A) not everything is being done in PID 1, and B) that most any init system is going to be to some extent or other a single point of failure. Now, generally, if you're managing a single system, the goal is to just make that single point of failure affect the least amount possible on the rest of the system, but the rationale with systemd's complexity seems to be making it easier to manage multiple systems. Which I would say is GREAT...if you are trying to rangle multiple systems as part of your use case (ie, massive VM farms). If, otoh, you're a home user, the benefit isn't there, while the cost remains.
Systemd isn't in itself bad...it's just not the best option for everyone's use case (and is demonstrably a horrible option in some). The issue here is that the "option" part is being threatened. Not something I'd care to support, though I assume Redhat already knew it wasn't getting a dime from me in the first place. Only recourse left then is to either A) actively develop alternative software, and B) aid in the proliferation of alternatives by educating the what and how of them. The real victims here are the new Linux users, stumbling out of Windows 8 into a battlefield of complexity, with fixes for various issues changing roughly every 6 months while entire systems are getting overhauled. Go ahead, try to use Google to troubleshoot common boot time issues with Ubuntu and tell me how clear and user friendly it all seems (and no cheating by specifying a specific init system...put yourself in a noob's shoes). The big distros apparently learned nothing from the disaster that happened as result of rolling out Pulseaudio in horribly misconfigured and unstable form before it was ready, and are repeating the very same mistake again. That Debian now is looking at possibly letting that kind of behavior into Stable (!!!) is really a bit of cause for alarm, and really will break quite a few jokes in the future (such as: Your mother is so old, Debian put her in the stable repository!).
(Score: 1) by ansak on Sunday October 26 2014, @05:02AM
Oh well, here's the best of it all: [debian.org]
The poster is wrong (in my insufficiently humble opinion) on plenty of things. But he's right on this: That's a solution I could be happy with. If certain RedHat developers who shall not be named in this point had started their efforts by offering this, I think there would be a lot less animosity expressed toward them.
But who am I to talk? ...ank
still looking for Gentle Treatment