VSIDO systemd and upstart discussion

[hakerdefo]

Thanks for that hakerdefo. Unlike gnome back then that had choices like xfce and kde and several other DE/WM's, systemd has basically upstart as an alternative and I do not think there is a question that sysvinit is lacking and needs to be replaced.  I am not sure if systemd or upstart would be the moving target in that analogy.

I wanted to convey two things by quoting Patrik. First, "If it ain't broke, don't fix it". Why fix-replace sysvinit?
Second, Patrik saw where gnome was headed before anyone else and gnome 3 proved he was right. So if he is skeptical regarding systemd then it is not a good sign.
And last but not the least good old Lennart Poettering is the brain(?!?) behind systemd.
Cheers.

[statmonkey]

hackerdefo, thanks for posting Patrick's comment.  I have read that article and the primary thread (with Lennart - I think) posing as a poster on LQ.  I think I posted this elsewhere but really do believe that anyone who is concerned with this subject should read http://0pointer.de/blog/projects/the-biggest-myths.html which is Lennart in his own words.  I won't say whether I think it will turn you on or off of systemd, just it will give you a good info base if you haven't read it.

Patrick also wrote this in regards to why replace sysv:

Well, the problem with using shell scripts in the boot process is that it goes through a lot of PIDs, and it would be "less ugly" to arrive at a usable machine state with a PID in the hundreds, or lower. If everything has to break in order to achieve that, it seems like a good trade. Eventually all the broken stuff will be fixed, right?

back in 2011.  Admittedly, he is the lone dev of a (not being critical) luddite oriented distro.  He has chosen not to add quite a few things and unfortunately I think Slack has suffered for that and fallen off the radar a bit.  He got some right, to give him his due (like HAL for example) but he has missed more.  Sorry but I am not going back to Slackware.  That thread http://www.linuxquestions.org/questions/slackware-14/slackware-and-systemd-885228/ devolved into a nightmare, some decent points were raised but it really does seem that there are a lot of arguments against Lennart and not systemd even Patrick posted:

Let's just switch to any crazy idea Lennart can come up with.

that quite frankly concerns me. Objectively, while I think Patrick is a Linux King the body of evidence would suggest that he trends toward the "conservative/no change/fine the way it is" line of thinking.  Again, not criticism just and observation.

I am undecided regarding systemd.  I want to be clear on that.  I see some things I like and really have some questions as to the impact of some of the decisions made.  I think it needs more time before I am ready to jump on the bandwagon (losing cron?; udev issues; not being able to grep logs;?) and I don't want to be posting novel sized comments on here all the time.  But I do think there is enough merit, enough positives regarding systemd and the problems it is trying to solve that it deserves fair consideration.  Certainly more than "the only thing it adds is fast boot times" which I am sick of hearing.  Or that this is an evil cabal designed by Lennart to take over the Linux world.  What I would like to have is more honest user experience.  I can't do this or that or it really helps this or I like that.

In the LQ thread linked above a member (who may or may not be Lennart ) posted the following in regards to Slackware:

It's customizable not because any effort was put into that. But because (contrarily to Debian and Fedora) it's one of the rare distros that does have a sane policy of not including everything and the kitchen sink. It makes choices instead of allowing every esoteric possibility. That's what makes it easily customizable. And it has to be this way. If not, one person wouldn't be enough to keep it updated. But the best would be a distro that would make choices, that doesn't include everything and the kitchen sink, but which would also strive to be consistent so as to put some order into the chaos of the bazaar. There's no reason why one can't have the best of both worlds.

I am guessing I don't need to explain why I like the quote and how it relates to this discussion.  My apologies as always for the long post and just hoping to add some value to the discussion.

[statmonkey]

I wanted to convey two things by quoting Patrik. First, "If it ain't broke, don't fix it". Why fix-replace sysvinit?
Second, Patrik saw where gnome was headed before anyone else and gnome 3 proved he was right. So if he is skeptical regarding systemd then it is not a good sign.
And last but not the least good old Lennart Poettering is the brain(?!?) behind systemd.
Cheers.

Oops! He certainly did get Gnome3 right and he may well be right about systemd.
You can read my earlier thoughts on Lennart but ...... while it's hard to get warm and fuzzy with the guy he ain't an idiot.  He does make it hard to leave the personal side out doesn't he? 

I think what you added was good input.  Again thanks for posting it, we post to generate thought and comment and certainly your post added value.  I am in hopes that all this moves our community to be better informed and prepared regardless of what happens with Debian.  Again, I am undecided on systemd and right now firmly in the if it ain't broke camp.  Currently reading all I can on runit and waiting patiently for the development and release of hakerdefo-ger.sysinit version 1.0 that fixes all of this 

[VastOne]

The elephant in the room is that sysvinit may not be broke, but is not doing what could/should/needs to be done

From the Debian Wiki

We've gotten a lot of mileage out of sysvinit in Debian, but its limits have been showing for a while - indeed, it was these limits that led to upstart being written in the first place.

   

• Sysvinit lacks service supervision. While /etc/inittab provides this capability, management of /etc/inittab is quite restrictive. Upstart eliminates the need for complicated PID file handling for all services. There are bolt-ons that allow you to do service supervision on top of sysvinit, such as runit, but it's clear that these are bolt-ons.


• Sysvinit does not track dependencies between services. Insserv/startpar provides this on top of sysvinit, but this is again very much a bolt-on, and only handles dependencies at boot/shutdown time (i.e., during runlevel changes) and can't handle any complicated service interdependencies at runtime. Upstart expresses this information in the native service configuration format, clearly and concisely.
• Sysvinit requires complex shell scripts for each service. While some of the complexity has been abstracted out into common helpers (lsb-functions; start-stop-daemon), having to represent each service's start/stop handling as a program is a severe handicap. Upstart jobs are simple, declarative, easy to maintain, and easy to modify locally as needed. They also eliminate the longstanding nuisance of update-rc.d reactivating services behind the administrator's back on upgrade.
• Sysvinit is linear. It stopped being a good fit for boot management on Debian the moment Debian adopted udev. There are many race conditions that persist in Debian today when booting with sysvinit, and although these may be fixable, the complexity for fixing them with sysvinit is very high. We're better off switching to an init system that's
  designed to work together with the event-based kernel and udev.

If this was not the case, we would not be anywhere near the angst all of this has caused and we would never have a conversation about it from debian's hard-line stance of "we won't fix it if it ain't broke'

IMO, this is more about change than anything... sure Lennart is there and Canonical is the other side and as Digit pointed out even more choices... What is telling to me is that there seems to be a hell of lot of distro's that are using systemd and that is documented... Which is not the same with upstart AFAIK, especially since there was a claim that it is being used in 'millions of servers, cloud based, etc etc'... This may be true but I have yet to hear it verified

Change is coming, whether we lemmings like it or not... and when the dust settles, I do not think sysvinit will even be a choice anymore because of it's shortcomings

I am very proud to be a part of the VSIDO community, you guys shine above anything I have ever been a part of

[statmonkey]

Nice post v-ger gives some of the _upstart plusses which have been hard to find.  Key point:

Change is coming, whether we lemmings like it or not... and when the dust settles, I do not think sysvinit will even be a choice anymore because of it's shortcomings

is, I think a fact which your version is about 75 lines shorter than what I tried to say earlier.

One thing I am concerned about and still researching.  It seems that since udev is wrapped in a module in systemd it might be necessary to have a separate udev outside that module.  The issue being one of control and flexibility.  I believe that IG mentioned this as well. Meant to include that in what I posted earlier but forgot.

[jeffreyC]

With respect to cron being eliminated by systemd, there are systemd services that replace at least some cron functions.

Some display managers do (or did not recently) work properly with systemd, the only one that it seems the systemd devs test against is GDM3 (which depends on most of Gnome 3)

I switched from SLiM to lightdm because at the time SLiM was not being maintained and was not working, that has changed and switching back would not upset me.

I am still on the fence about systemd, I see valid arguments on both sides.

[statmonkey]

JeffreyC

Some display managers do (or did not recently) work properly with systemd, the only one that it seems the systemd devs test against is GDM3 (which depends on most of Gnome 3)

Add this to the concern list from two angles.  One that they do seem too interested only in what they see as "necessary" items and it is a short list.  Their Gnome lust and the fact (for example) that they are pretty much saying "BSD users aren't important screw-em". It's easy as a Debian user to get lost and forget about other 'nixes but history shows the group that is getting told to eat dirt today could be the group you are involved in tomorrow.  I don't like the idea that somebody (particularly Red Hat/Lennart Poettering) are telling me what is of value and what isn't.  This is not the Unix/Linux philosophy or attitude.

FWIW I haven't had a problem with slim or lightdm under systemd. 

In summary my concerns are (not in order):

• Red Hat
  Lennart
  Bundling of core systems
  What will be removed and bundled in the future
  Access to udev/init level functions for api purposes
  Flexibility going forward (software choices, configurations, etc.)
  Dependencies/for example Gnome depending on systemd -- bad trend
  The dictatorial attitude of what is valuable and what isn't
  What if something they don't think is valuable but I do is dropped can I have access/ability to build and run it at init?  At some level?
  
  The fact that I can't find any user statements by unbiased parties giving me facts as to why they love systemd. I can find new users who post about it but overall very little about plus or minus. Does this mean it works so well there is nothing to say or does it mean it does so little there is nothing to say? Where are the people who installed it and hated it?  Loved it?

I really need more hands on experience with systemd and the nuts and bolts of it.  I (as of now) haven't found anything that I can't do and the bugs that I have found are livable.  Several people have posted that they have problems but I have yet seen anything that isn't or could not be fixed.  My position remains that I have a small sample size of real evidence in favor or against.  Some people I respect endorse it, some people I respect are strongly against it.  This goes for some distro's, it does mean something that Arch and Siduction have embraced it.  About the only conclusion I have come to is one I had a long time ago and that is that sysv is inefficient and should be updated.  Not exactly Einsteinian analysis on my part.

Obviously this discussion is a hot button for me.  I think the last time anything was this polarizing was Gnome 3 and we all know how that turned out.  This feels bigger and more impacting to me but I may just be overstating the case due to ignorance.  It's that ignorance that really makes me uncomfortable.

[jedi]

@statmonkey
You should indeed be skeptical.  This is my second go 'round with systemd.  Not that he would remember little ole me, but i have corresponded a couple of times with Lennart.  Both times he was kind and forthcoming.  On the other hand, I've read posts where it almost seemed like he was ready to beat someone to death!  No where in my Linux travels have I ever found a better, more knowledgeable community than right here at VSIDO.  As well as being kind and totally human in their communications!

This time running systemd has been much more successful to me.  I think the cron/grep issue is moot.  The journalctl abilities are getting better everytime I use it.  I believe this time I will finally make the decision to stick with systemd.  I am in no way trying to influence a decision on anyone reading this threads posts, I'm just writing about how things are working out for me.  All of the /var/log files are still there and astutely collecting all the log data they should be.  You CAN still use grep with journalctl.  There is a small bit of reading, or a small learning curve involved.  When new tech comes along, when is this not true?  It is quite painless, and if I can muddle through, anyone can.

To revisit Upstart, the mess with the copyright stuff scares me.  This is totally against my studied philosophy regarding FOSS and OpenSource.  Also Canonical/Ubuntu purposefully limits the amount of dev's allowed to work on the Upstart project.  The opposite of this is true with systemd.  Also with systemd, Lennart is just one of many cogs in the wheel of its development.

Since April of 2012 when udev was merged with the systemd project, the dev's with systemd have worked hard on the upstream maintenance of udev, and indeed there are still udev dev's working with them.

I happen to love sysVinit.  Have used it for years.  Will hate to see it go.  That said, it will go.     My feelings are, we are better prepared when that day comes, if we already have a plan in place to replace it.  Though the VSIDO alpha ISO I now have installed is an 'alpha' release, it runs as if it were an officially released VSIDO distro.  This says tons about our dev, VastOne.  As a matter of fact, I decided this morning to make it my main goto OS.  Of all the distro's I've tested and used through the years, this one stands out the most because of ONE dev's dedication and willingness to work with a community of users.  He listens!

As to the previous post about Slackware, you had some outstanding insights as to it's developer.  He is a King of Linux, there is no doubt.  But the lack of willingness to try new things over there is a big turnoff for me.  You summed it up perfectly with the term luddite.  For those reading this, that don't know what that means, google is your friend!  I mean absolutely no offense in this regard either!

When it comes to systemd, I believe it is a larger, more open minded, and more dedicated group of dev's than a lot of people want naysayers to believe.  It is totally under the GPL unlike Upstart which is covered by copyright.  It has a much larger team (systemd) that is willing and has proven their dedication to Debian.  I truly fear that Upstart would hurt Debian in the long run, and do not have this fear with systemd.  Reading up on Linus Torvalds calling out Kay Siever (one of the udev dev's) about submitted bugs and the patch made available shortly thereafter was eye opening.  If nothing else it is amusing and shows that Torvalds is still paying attention to all things Linux.  (http://article.gmane.org/gmane.linux.kernel/1369384)  I'm sure we're all aware of Torvalds lack of empathy regarding stupidity and or willful ignorance.  He calls it like he sees it and the link above will give you a chuckle.

All of this said, until absolutely necessary, I see no reason to abandon sysVinit at this point.  One benefit to changing right now however, would be to end up on that list of growing distro's adopting systemd as it's default.  I believe firmly that Debian will certainly take that list into account when making their final decision.

One more thing, I think systemd has been working hard to make the cgroups functionality better and better, as well as making udev work better as well.  I would think if it is finally adopted as the default that the module type of operation of udev outside of systemd will be unnecessary.

Just my empty headed observations on this and please, take it all with a grain of salt.  (or a shot of bourbon)   

[statmonkey]

Jedi, tremendous post.  Just what I was looking for.  Your opinion matters very much as you are certainly among those Vsido members that make the community so strong, also because over my time here I have learned to respect and trust your opinion and insight.  That is no small compliment, I am by my nature resistant to giving out said trust and respect.  I do not think it is coincidental that there are several people in these forums whose opinions I have come to value.

This time running systemd has been much more successful to me.  I think the cron/grep issue is moot.  The journalctl abilities are getting better everytime I use it.  I believe this time I will finally make the decision to stick with systemd.  I am in no way trying to influence a decision on anyone reading this threads posts, I'm just writing about how things are working out for me.  All of the /etc/var/log files are still there and astutely collecting all the log data they should be.  You CAN still use grep with journalctl.  There is a small bit of reading, or a small learning curve involved.

Extremely valuable input.  This is exactly the type of comment I have been hoping to generate.  Even if I think you are undervaluing your knowledge.  There is more in that paragraph and post as a whole regarding actual usage of systemd than I have seen anywhere despite reading hundreds of posts on the subject.  In fact there is probably more in this thread than I have seen anywhere.

I am sure I am not alone in saying that your post has influenced me and pushed me a little harder to remove my ignorance and dive in.  I am posting this from a systemd box and have my laptop also configured with alpha so at least for now I am in a position to relieve some of my skepticism or confirm it. 

As for _upstart I have already thrown that in the no go pile.  Many reasons that have already been posted including you insights.

Thanks for getting me to re-read the infamous Sievers post.  It did put a hop in my step.  Good (insert your favorite supreme being here) if I got an email like that I would slink back to my cave and never come out. 

Your post is hardly hair-brained and while I am not a bourbon drinker I will have a little Glenfiddich in your honor.

[lwfitz]

Wow, I stay away for 6 months and look what I walk in on....... Some seriously good posts in here, time for me to catch up........

[Digit]

statmonkey said:
Digit - you actually are exactly the one who should be testing this.  [...snip...]  I understand it being a pain in the ass but someone who uses a lot of software is probably going to see different things than those of us who run somewhat minimal setups.

yup.  i totally get that... but given it's the production machine i run "always on", and depend on not only as my heavy workload workstation, but also my server to feed all my other machines...  i'm sure you can appreciate my reluctance to rock its boat. 

... still, that's all rather mute now.  it seemed to have a catastrophic hardware failure last night.  :/  ... systemd was getting pulled in and updated already... idk if that's related.  lol.   
(side note... pcpartpicker is a flippin awesome site)

various peeps mentioned:
[...snip...]slackware[...snip...]

YUS!  gotta admire the strong steady rock of the unwavering philosophy.  may not appeal to the magpies wanting to jump on every new bandwaggon, but i really admire and am maaaasssively grateful for slackware's persistence as it is, not changing that which isnt breaking, and, as was strongly hinted at earlier, volkerding paying attention and moving out of the way of falling trees. 



besides talk of how foss each of systemd and upstart are, there's something that's been lurking between the lines a lot here, that being, how "unix philosophy" they are.  i mentioned runit earlier...  it's likely one of the most "unix philosophy" of the lot, but then, i gather from a lot of the qualms folks have with sysvinit, it might be covering some of the same ground.  to my way of thinking, the unix philosophy is at the core of what makes our much loved GNU|Linux systems so potent.  much as one might say all that GNU needed was a kernel, all that unix needed, was freedom.  which linux, and gnu gave, respectively.  again, my disclaimer that i dont know enough to say anything for certain, but from the sounds of it, BOTH systemd and upstart piss on the unix philosophy to some degree.  i think that's what makes me most uncomfortable with them.
... the thought of gaining some, but losing a lot.
(i've got an analogy on hand for pharmaceutical medication and side effects, but i cant seem to word it right without muddying my point)


as for the foss'ness of upstart...  it does seem rather weird.  http://upstart.ubuntu.com/wiki/CopyrightAssignment  what?  ...  ... why did they even bother?  what are they up to?  ... something fishy.   almost as if it's foss, but a corruption of foss with their own backdoor paranoid powergrab.  *headspin*
GNU GPL licensing depends on copyright legal framework too... it just tends not to do so like anyone wants to retain the power, and to have copyright of the contributors muted by passing on to, in this case, canonical.  o_O  even when fsf or gnu take on board consolidated copyrights, they tend not to do it in this way.  idk... maybe i've just had my thoughts tainted about it with ominous suggestions.


anyhoo... there's still more reading on this subject i want to do (particularly to find out more about the results from those who managed to get openrc in debian working).  this thread's barreling ahead of my researchings. 


oh, one other thing...  to add to the choir of collective self adulation & admiration amongst the vsidoans,
a discordant note if i may...
stfu, u all suck, disapearing up yer own arses!

hehe.  just wanna keep us on our toes, n not get too full (or fool) of ourselves.    much love.
(but yeah... tis why i am here... "the only way to get better is by playing a smarter opponent" -revolver (2005) (my fave film))


*gone to read more*
oh, as i hit post, it tells me i might wanna check the two new posts... yup:

pushed me a little harder to remove my ignorance and dive in

this thread sure excels at that.

[edit] also re-skimmed parts of this thread... not sure if i'm adding anything new, just restatements that are getting closer to throwing a tantrum [/edit]

[hakerdefo]

Here is a story my ma used to tell me when I was a child,

A quail had her nest in a wheat field. She had five babies in the nest. All the little quails were too young to fly. One day the farmer and his wife came to the farm and ma quail and her babies overheard the conversion.
Farmer's wife -> "I think it's time we start harvesting."
Farmer -> "Okay. I'll hire some farmhands and begin harvesting next week."
All baby quails began to cry. Asking the mom "What will happen to us? We can't even fly? And you ma told us it will at least take another month for us to be strong enough to fly."
Ma quail replied "No worries my children. We all will be fine."
The farmer returned next week with his wife.
Wife -> "When on earth are you planning to begin the harvesting?"
Farmer -> "All the farmhands are busy harvesting in the other fields. I can't do it on my own. I'll start as soon as the farmhands are available."
On hearing this baby quails again started crying. Ma quail said "Listen my children. I know what I'm doing. Have no worries. We'll be fine." All baby quails had their doubts but there was nothing they could do.
Much to their relief, joy and surprise a month passed by, They all learned to fly, Neither farmer nor his farmhands came. They asked ma quail how had she known that farmer wouldn't be there to harvest the wheat crop. Ma quail replied,
"Listen my young ones. The person that relies on the help of others doesn't have a choice. He becomes dependent. Dependency brings laziness. Lazy and dependent person loses the control over his own affairs and life. So my darlings you all should learn from this and always strive to become self-reliant and self-sufficient. "

No matter what Debian chooses, It's gonna rely on others.
Hell with upstart, Hell with systemd. Debian, build your own new init. It will take time, effort, blood, sweat, tears. There are people out there who are and who will be ready to give all the required effort-blood-sweat-tears. How long Will it take? A year, May be two at max. For sure sysvinit can be made to serve Debian for another year or two. Yes it might be tempting and attractive at this stage to go for ready-made solutions and spare all the effort-blood-sweat-tears. But we all know that Beggars Can Never Be Choosers.
Cheers.

[VastOne]

^ That... Great story and then great direction

Well done!

@Digit and Jedi, incredible feedback mates, well done!

[Digit]

Hakerdefo said:
Here is a story my ma used to tell me when I was a child,

A quail had her nest in a wheat field. She had five babies in the nest. All the little quails were too young to fly. One day the farmer and his wife came to the farm and ma quail and her babies overheard the conversion.
Farmer's wife -> "I think it's time we start harvesting."
Farmer -> "Okay. I'll hire some farmhands and begin harvesting next week."
All baby quails began to cry. Asking the mom "What will happen to us? We can't even fly? And you ma told us it will at least take another month for us to be strong enough to fly."
Ma quail replied "No worries my children. We all will be fine."
The farmer returned next week with his wife.
Wife -> "When on earth are you planning to begin the harvesting?"
Farmer -> "All the farmhands are busy harvesting in the other fields. I can't do it on my own. I'll start as soon as the farmhands are available."
On hearing this baby quails again started crying. Ma quail said "Listen my children. I know what I'm doing. Have no worries. We'll be fine." All baby quails had their doubts but there was nothing they could do.
Much to their relief, joy and surprise a month passed by, They all learned to fly, Neither farmer nor his farmhands came. They asked ma quail how had she known that farmer wouldn't be there to harvest the wheat crop. Ma quail replied,
"Listen my young ones. The person that relies on the help of others doesn't have a choice. He becomes dependent. Dependency brings laziness. Lazy and dependent person loses the control over his own affairs and life. So my darlings you all should learn from this and always strive to become self-reliant and self-sufficient. "

No matter what Debian chooses, It's gonna rely on others.
Hell with upstart, Hell with systemd. Debian, build your own new init. It will take time, effort, blood, sweat, tears. There are people out there who are and who will be ready to give all the required effort-blood-sweat-tears. How long Will it take? A year, May be two at max. For sure sysvinit can be made to serve Debian for another year or two. Yes it might be tempting and attractive at this stage to go for ready-made solutions and spare all the effort-blood-sweat-tears. But we all know that Beggars Can Never Be Choosers.
Cheers.

yus.  well illustrated hakerdefo.

it surely cant be THAT hard.  if a pokey lil distro like dragora can seemingly jump to it without any fuss...
... oh, i wonder if that's part of the problem...  debian lost any chance of a nimble development that can turn on a penny.  ^_^  i hope that shames someone* into trying.  lol.

surely there could be more blue-sky wild-notion research-developments...

like was said in one of my fave foss related talks (and from an ubuntu/debian guy no less), failure is cheap & success is cheap.
[edit] to which, i'll add:  change is cheap[/edit]
... not to mention surely a lot has been (or at least can be) learned from systemd's savage thrust into the heart of GNU|Linux, meddling with important core components, giving things a much overdue radical shake-up.


* no one specific in mind... but if you thought i meant you, then sure, yes, i meant you. 

ps, after half a day of reading through debian mailing list archives on this topic... i see what you guys mean about quality of input.  ^_^  ...makes me feel better about my own poorly informed drivel.  lol

[statmonkey]

No matter what Debian chooses, It's gonna rely on others. ...

Hell with upstart, Hell with systemd. Debian, build your own new init. . But we all know that Beggars Can Never Be Choosers.
Cheers.

Digit re-iterates the point.

This seems at cross purposes to me.  Debian is not going to rebuild an init or to put it bluntly choose the path of most resistance.  To paraphrase v-ger "The Elephant in the room here ..." is that the clearest path if we reject _upstart and reject systemd is to develop our own. 

Digit

it surely cant be THAT hard.  if a pokey lil distro like dragora can seemingly jump to it without any fuss...

Well, they essentially didn't that really looks like a repackaging of sysv to me.  But the point remains that:

But we all know that Beggars Can Never Be Choosers.

So what do we have?  A choice between accepting the status quo or rising up as a community to move toward our own solution.  Besides impetus what do we need? 

• A real core level understanding of the problems of sysv
  A real core understanding of  the alternatives
  Time and effort and co-ordination
  A skill set to develop it (based on the multitude of skill sets here)

None of these things seem lacking here.  Is it realistic to think we can/would do it?  I don't know, I certainly am not (at this moment at least) knowledgeable enough on any of these subjects but it seems clear that the following is where we are:

• Wait for Debian to decide something and accept it
  Wait for some miracle to happen and somebody else to do something
  Embrace systemd and move on
  Embrace _upstart and move on
  Accept one of the janky alternatives
  Alter one of the janky alternatves
  Create our own

Not deciding is a decision (the first one in this case).  Depending on others is waiting for the farm hands.  I am not calling anyone out here, I am just stating the case as I see it.