VSIDO systemd and upstart discussion

Digit

easy as piss.  no worries then. 

:) 

woohoo, systemd ootb in vsido!

*converted*

statmonkey

Quote from: Digit on January 04, 2014, 11:21:48 PM
*sniff sob bleet*

my remaining concern that's not just probable fud, is that this contributes to a sort of intelectual entropy in the broader exosphere... in that this helps solidify the compatibility separation from non linux unixes just a little more.  yep, it may well indeed be good for us...  we shall see...  but i think systemd addoption may be a sad note for broader foss.


My fear as well as I stated earlier. 

BTW I have an existing VSIDO install that isn't systemd is it the reverse of the above to add it or is there more?

VastOne

VSIDO      VSIDO Change Blog    

    I dev VSIDO


hakerdefo

Here are the people on the technical committee that is going to decide on the default init system for Debian,

Andreas Barth
Bdale Garbee
Colin Watson
Don Armstrong
Ian Jackson
Keith Packard
Russ Allbery
Steve Langasek
Now I do not question or doubt integrity of the members but I'm gonna make some assumptions.
Ian Jackson, Colin Watson and Steve Langasek were or are employees of  Canonical. So I'm assuming three votes going in the favor of upstart.
Don Armstrong in his recent comments has indicated that he doesn't like the complexity of systemd.
Bdale Garbee is highly unlikely to go for systemd.
Andreas Barth and Russ Allbery are going to go for systemd if we go by their recent posts and comments.
Keith Packard has his cards closed to his chest so the direction of his vote is unknown.
So if we conclude by the above assumptions upstart is going win by the votes of 5 to 3 or 6 to 2.
Cheers.
You Can't Always Git What You Want

PackRat

^ That's funny in an ironic sort of way. Most any other situation, the 3 that are also employees of Caononical would have to recuse themselves from voting because of conflict of interest.

@vastone - should you decide to go forward with systemd, are all the system admin apps installed by default (bum comes to mind) systemd compatible?
I am tired of talk that comes to nothing.
-- Chief Joseph

...the sun, the darkness, the winds are all listening to what we have to say.
-- Geronimo

VastOne

^ I agree, having those 3 involved in the decision seems a bit unfair, which is what jedi has been saying all along that scares him.  I like to think that decisions will be based on technical merit alone, but that seems doubtful

I have tested every part of what is stock in VSIDO and all the build tools I use including VirtualBox, debian-live etc etc and have not found anything that will not work with systemd or that has any negative impact at all
VSIDO      VSIDO Change Blog    

    I dev VSIDO

statmonkey

#82
Quote from: PackRat on January 05, 2014, 05:02:06 PM
^ That's funny in an ironic sort of way. Most any other situation, the 3 that are also employees of Caononical would have to recuse themselves from voting because of conflict of interest.
The other night I was discussing this very thing with someone.  I would say it is more than funny and ironic, it's dishonorable.  There I said it.  If you have a vested interest in a program (in this case) honor would compel me and precedent would imply that you would either abstain or excuse yourself.  Like I accused someone else of, I guess I am a Luddite on things of this nature or I am just weird.  Thank you though PackRat for saying that it makes me feel that I am not a lone man in the wilderness.

I really do think Debian will go the Canonical Way and posted my reasons earlier.  (shudder) The hanging chad in this situation remains the political implications not the technical side of systemd.  I can't find anything that I can't make work or doesn't just work.  But the philosophical issues:

systemd not being portable
Debian's position/future behavior

are all that I am really bothered by.

If VastOne goes in this direction does what Debian does really matter?

LOL I keep saying to myself stfu on systemd but then someone posts something I can't help but respond to.

[edit] Maybe we are missing something obvious.  Digit and hakerdefo have both posted arguments in favor of looking at alternatives to systemd and not accepting the status quo because it is there.  Perhaps we are looking at this from the wrong level.  Given that you have Arch, Fedora, etc. who have already accepted systemd (I know of a couple of other Debian distro's that are seriously considering it but have promised not to say more) the status quo that won't be accepted here is not a re-invention of the init level and a better solution but the re-invention of Debian itself or the creation of an alternative.  I am not sure I have ever felt that such a thing was really possible before.[\edit]


VastOne

Quote from: statmonkey on January 05, 2014, 06:08:23 PM

If VastOne goes in this direction does what Debian does really matter?

That is the problem I am facing though.  I want to  remain as close to the core of Debian as possible and obviously  init is part of that core

I am also curious as to how this will fallout, no matter what the decision is.  Debian's track record on releases that center around stable is well documented.  Will the decision be made in time to change over in this cycle of stable, no matter which it is?

I think what I will do at a minimum is wait for the decision to be made from Debian, which has already supposed to have happened.  Once that is done, and the decision is for upstart, I would think a complete rebuild of VSIDO with the core that includes upstart will be built and tested. 

I will also build a Alpha release now with upstart so that we can have a side by side comparison of systemd and upstart

Will let you all know here when it is done
VSIDO      VSIDO Change Blog    

    I dev VSIDO

PackRat

Quote from: statmonkey on January 05, 2014, 06:08:23 PM

I really do think Debian will go the Canonical Way and posted my reasons earlier.  (shudder) The hanging chad in this situation remains the political implications not the technical side of systemd.  I can't find anything that I can't make work or doesn't just work.  But the philosophical issues:


be interesting if some of the Debian dev's split off and start a new project/fork should upstart get adopted.
I am tired of talk that comes to nothing.
-- Chief Joseph

...the sun, the darkness, the winds are all listening to what we have to say.
-- Geronimo

statmonkey

Quote from: VastOne on January 05, 2014, 06:42:32 PM

That is the problem I am facing though.  I want to  remain as close to the core of Debian as possible and obviously  init is part of that core

At what price?  At any price?  Just asking.  Having a test version with upstart would be a good thing (if not too much work) but your research has led you to favor systemd.  I am struggling how to put this correctly.  There are times when a relationship is just grown out of.  I personally have a loyalty issue.  I am extremely loyal to people who do me a good turn, often to my own detriment.  The favor is paid back and I continue to feel I owe them something long after the ... balance in the relationship has gotten extremely one-sided.  It is a painful lesson I learn repeatedly so I am perhaps over-sensitive. 

The point I am attempting to make is that you have built a brilliant distro.  To my personal needs it is the best I have ever used and combines all I could want.  If I had the time or inclination to build my own I don't think I would do much different and could only hope to achieve half as much.  The distro seems to be predicated on the concept of the right tools/environment/configuration for a semi-technical user to boot and be comfortable immediately without all the cruft, retaining a low footprint and memory.  What goes on in the user space and below is what is best regardless of what others may promote but what works and is solid. All indications at this point lead me to believe that systemd meets both criteria and upstart does not. 

Early in this thread I warned that this is the type of topic that can become consuming (it has) and can easily make the best of us lose focus of the end-goal.  I don't presume to know your end goal, but that statement above sets off some alarms for me.  I trust your judgement fairly implicitly, your considered opinion is of high value but if there is indeed a fork in the road I have far more faith in your analysis and decision making skills than a committee of Debian ne'er-do-wells.

I hope the respect and admiration that I have is reflected in the above, I like Debian a great deal but if they accept upstart it's a no go for me and I will part ways.  The reasons have nothing to do with upstart and everything to do with upstart.  The jig, as they say, would be up and Debian as we know it done and dusted.  This for me is a line in the sand, because once Canonical gets their hooks set into Debian it's over.   Perhaps this is an over-reaction but I doubt it.   :( 

Yes, I am an annoying pain in the arse.

VastOne

I like annoying PITA's...  ;D

I think you misunderstood my intent, or I did not speak it clearly.  Want and need are the questions at hand.  As much as I want to stay at the core of Debian, I do not need to and obviously the community will make that decision

I have questions though on what you have said... Is your dislike of upstart purely philosophical?  Is it just because of Canonicals involvement?  The Canonical license?  Debian simply not existing anymore if they adopt upstart?   If upstart was the best thing ever and blew anything else away, would you still feel the same?

I watched this video from DebConf13 Why Debian needs Upstart
I was happy to see the angst at the end of the presentation by the questions raised by Ian Jackson who was clearly frustrated.  I was also shocked at how flippant Steve Langasek was about the license issues Ian was raising.  I think I understand now why the decision is taking a long time from the technical committee

This is a huge deal for Debian as we know it and your concerns are valid.  I did not mean to imply I had changed my opinion on the matter, I was only suggesting that we should take a look at both to be sure due diligence has been done

Pottering was also at that DebConf13 presentation and had some questions and also gave his reasons why systemd should be used also ... you can see that on one of the links from that video
VSIDO      VSIDO Change Blog    

    I dev VSIDO

VastOne

Here is a pretty good video that shows the differences between sysVinit, upstart and systemd at a functional level

Init, Systemd, and Upstart - CompTIA Linux+ LX0-101, LPIC-1: 101.1
VSIDO      VSIDO Change Blog    

    I dev VSIDO

statmonkey

#88
Quote from: VastOne on January 05, 2014, 07:52:13 PM
I like annoying PITA's...  ;D

I think you misunderstood my intent, or I did not speak it clearly.  Want and need are the questions at hand.  As much as I want to stay at the core of Debian, I do not need to and obviously the community will make that decision

I have questions though on what you have said... Is your dislike of upstart purely philosophical?  Is it just because of Canonicals involvement?  The Canonical license?  Debian simply not existing anymore if they adopt upstart?   If upstart was the best thing ever and blew anything else away, would you still feel the same?

I was happy to see the angst at the end of the presentation by the questions raised by Ian Jackson who was clearly frustrated.  I was also shocked at how flippant Steve Langasek was about the license issues Ian was raising.  I think I understand now why the decision is taking a long time from the technical committee

This is a huge deal for Debian as we know it and your concerns are valid.  I did not mean to imply I had changed my opinion on the matter, I was only suggesting that we should take a look at both to be sure due diligence has been done

Pottering was also at that DebConf13 presentation and had some questions and also gave his reasons why systemd should be used also ... you can see that on one of the links from that video

[edit]Just watched Lennart his social skills are limited aren't they.  But he does come off like a human versus the robots from Canonical.

My problem with upstart is also technical.  In summary:

Systemd is a complete rethink and rebuild of how the init system is conceptualized and makes full use of modern systems.

Upstart seems a slapdash attempt to get control of the init process but at the same time not really alter the issues that are there in sysV.  It's kind of like the first Windows GUI.  Just creating an icon and a batch script and calling it a graphical user interface doesn't make it one.  Just adding a layer on sysV and calling it a new init system doesn't really make it one.  Finally socket handling or lack there of.[/edit]

Just watched the upstart one and will watch Lennart's show and the new link.  Thank you.  I think my thoughts are confirmed by two things.  One is at around the 16:20 mark where Steve essentially says "it won't work because it's upstart dependent and not doing it with upstart makes it fail" he re-iterates the same idea with the GPG question at the end.  You have systemd clearly admitting that there are going to be dependencies with systemd while the upstart people are saying no we are not causing dependencies but if you don't write it specifically dependent on upstart it won't work.  Like the "corporations should be allowed to control and have the license for what you write" statement, I can't accept that as a foregone conclusion and the insidous-ness (word? I think not) of the smarmy Steve and his glossing over made me throw up in my mouth.  The guys question on the event structure is legitimate and laughed off.  I guess this only confirms my feeling that today the battle is over upstart but the war is over control of Debian.  Once you bend over the first time .... sorry that is a little graphic.  In answer to your questions.

My concern is over the license - first and foremost to agree to that gives Canonical a theoretical ownership of a core process in Debian.  That foothold at the base is a predication to control.  Power is never given it is taken and Shuttlegates is who he is.  Having control of Debian's user base and the ability to add those installed systems as "yours" is worth $MILLIONS to Canonical.  Don't kid ourselves this is what it is about.

My concern is over the control of Debian and it's direction.  Listen to what is said, it's already Ubuntu this and Ubuntu that and Debian is thrown in as an after thought.  Already a step-child without the formal adoption.

My belief is that once Canonical gets control of the init process the ability of ANYONE to stop them from the Ubuntu-ization of the whole Debian is gone.  We have to have Unity because it works so well with Upstart, etc.

LOL at Steve and his "I don't understand why the license is so important to Debian people".  I will with-hold further comment due to this being a family site :)

Bear with me for a personal aside.  At one time in a galaxy far away I was involved in corporate takeovers.  One of the strategies we used quite successfully was the takeover by proxy of boards and decision making.  I often got companies without ever having to pay anything by just controlling decisions and influence.  Once a certain momentum was reached I never had to ask but just say, well we already did this and that so this is just the next natural step.  The former owners never knew what hit them.  There is an eerie similarity here and I think that is what has my spider-sense tingling.

zbreaker

statmonkey, you hit the nail on the head with the above thoughts. I've gotten a sick feeling in the pit of my stomach as to the possible consequences of Debian adopting upstart. I can only hope I'm dead wrong if it does happen.