How to install nosh init system on FreeBSD?

mark_j said:
Meh, parallel startup? I'm not attacking your 'nice to have' opinion but it's overrated.
Click to expand...
JFTR, I didn't say I want or need that ;) Only thing I was saying is that this is the only thing I could think of which *might* be improved in an init-system.
 
mjollnir said:
[...]
IMHO if an alternative init system is done right, it would neither hurt nor restrict you, but enable less skilled people to govern it.
Click to expand...

With respect that's phooey.
Your average user & even most kernel developers don't even touch their init system. On NetBSD we've had this same argument and it's (always) under misapprehension. If any one is impacted it's the porters, & believe me they don''t want to supports n-amount init systems.

Ibm/linux pushed this because there's hundreds of distributions that have disparate init systems that require greater work by distro developers and packagers; systemd init was born (and an ugly bastard it is!).

Contrarily, bsd has no such issue, thus rendering your argument moot.
 
mark_j said:
Your average user & even most kernel developers don't even touch their init system. On NetBSD we've had this same argument and it's (always) under misapprehension. If any one is impacted it's the porters, & believe me they don''t want to supports n-amount init systems.
Click to expand...
By n-amount do you mean parallel startup or disparate init systems? The latter would not happen since there's one FreeBSD. At least one FBSD-based distribution (TrueOS) used OpenRC IIRC, but TrueOS is history now.
Ibm/linux pushed this because there's hundreds of distributions that have disparate init systems that require greater work by distro developers and packagers; systemd init was born (and an ugly bastard it is!).
Click to expand...
See above: I'm not promoting systemd... In contrast, my concern is: it's reasonable to evalute runit vs. OpenRC as a replacement for the current BSD init, like e.g. is done here for Linux. Both allow for parallel services start, and they might be easier to manage for newbies (that's to be evaluated). Additionally, they supply service monitoring/supervision, which seems to be a demand of professional server plants. It should easily be possible to switch off parallel startup, I you want that. Likewise, it should be possible to not monitor a service. Both are reported to be fairly slim & KISS. Thus, you'll loose nothing, while others could benefit.
Contrarily, bsd has no such issue, thus rendering your argument moot.
Click to expand...
It does not, because the current BSD init does not have a feature that is demanded by many desktop users: parallel service startup. The only arguments against this so far were 1. added complexity, which I doubt; at least the added complexity will not be overly much IMHO; and 2. race conditions, i.o.w.: in most cases, it works well. There are race conditions in any system beyond a certain complexity, e.g. currently my wpa_supplicant(8) does not start automagically, but it does start w/o errors & warnings manually. If I want parallel startup, I would probably run into many issues right now. If it comes by default, I'd be happy.

Again: please stop complaining about systemd and using it's deficencies as contra argument. We are 100% d'accord that this mess is broken by design. Come up with arguments against runit (sysutils/runit) and/or OpenRC instead.
 
Well, OpenRC seems to work fine with a BSD init and still use init scripts, so this might be acceptable (while a solution trying to replace scripts with configuration files will never be acceptable to me). Still, why change something that isn't broken? So OpenRC can do parallel startups? Really, I don't need that. Everything I need is already present with mewburn rc and works reliably and proven for many years.
 
kpedersen said:
FreeBSD is a simple tool. Just like a hammer, it is impossible to make it without replacing it with something else entirely. If people want a "fun" experience and to play, they don't use a hammer and they use Microsoft Windows instead.
Click to expand...

Some people insist on making an operating system "user friendly". They care a lot on "less skilled users". Every operating system is getting such an operating system, and there is always fewer "simple tools" as you call FreeBSD. What will be at the end the alternative for the "more skilled users"?

I consider myself not a skilled user, that is why I have great problems with operating systems for less skilled users: you can use them, but when there are troubles, you cannot do anything, because they are too complicated, difficult to understand and to deal with, and troubles come always, because they are complicated.

If I wanted a OS for "less skilled users", I would use Ubuntu or Windows instead of FreeBSD and OpenBSD. I do not criticise Ubuntu or Windows, they have they user basis, they have a function, but please, do not take me my OS away!
 
mark_j said:
Meh, parallel startup? [...], it's really only ever going to add complexity & huge chance of race conditions
Click to expand...
All (?) the information needed for parallel startup is already present in the dependency hacks in the service scripts. Thus, it would very likely add very little complexity, maybe even none at all.
hruodr said:
[...], but please, do not take me my OS away!
Click to expand...
There's a good chance that both runit & OpenRC are KISS enough that you won't have to worry about this. In fact, IMHO FreeBSD is MUCH simpler than Windows & Ubuntu, if you're a little bit curious & courageous.
 
Zirias said:
This wording (configuration) brings me to the next criticism: shell scripts. With mewburn rc, there's a large and useful library, so your init script for a "normal" service needs only a few lines. Dependencies also work rock solid. Still, with these scripts, I can do anything unforeseen by the designers of the init system. I personally wouldn't ever want to use a system that replaces scripts by some "configuration".
Click to expand...
If the contents of the configuration is "use the default script", that would satisfy both, yours & my concerns, right?
 
mjollnir said:
I know many ordinary, non-nerd people who switched from Windows to Linux for various reasons. If they see FreeBSD starting up, they call it computer-stone age.
Click to expand...

True but these same people say that C is "old" compared to Java and have been for years. They would be wrong. Tools such as FreeBSD (and hammers) don't need pretty startup screens.

Non-nerd people honestly don't need real computers anymore. They need a telephone or a games console. It is impossible to cater for them without replacing FreeBSD with a different project entirely.

I would even consider myself a "non-nerd" user. I am not an OS developer. Heck, I am a slow tech learner which is probably why I think Linux is defective. I just use FreeBSD to provide a platform to develop software and teach on. Windows and Linux have so much breakage daily that they no longer provide a solution to my needs. How is that user-friendly?

mjollnir said:
Please come up with arguments against runit & OpenRC instead.
Click to expand...

The sad truth is that these haven't seen much uptake because they don't offer more than systemd or existing standard init systems. There are a number of arguments why they are not popular, just do a search for the "systemd vs openrc" threads. In terms of FreeBSD they do not offer enough to justify the breakage they would cause.
 
mjollnir said:
If the contents of the configuration is "use the default script", that would satisfy both, yours & my concerns, right?
Click to expand...
This might be the case indeed, but I'm still not convinced such a central part of the base system needs any change at all. Mewburn rc has (IMHO) a very elegant design and I personally don't miss anything. Changes are never free of risk (and FreeBSD tends to be much more aware of that than Linux), so to change something, you need good reasons. I see we just disagree about these "good reasons" ;)
 
All the above said, It would be great to see nosh (and all other init systems) in ports for those that really do have a use-case that can benefit from their features.

I expect zoujiaqing has a lot of work ahead of him :)
 
kpedersen said:
Non-nerd people honestly don't need real computers anymore. They need a telephone or a games console. It is impossible to cater for them without replacing FreeBSD with a different project entirely.
Click to expand...
They can decide for themselves... Ordinary people want a desktop computer system to write e-mail, chat, browse the internet, etc.pp. Note the act of writing. Other devices like tablets & smartphones, smart TV & game consoles do not have a keyboard and/or the display is too small. That's two of the reasons why laptops and mini computer desktop systems are still very popular. My assumption/assertion are:
  1. adding parallel service startup will not hurt you, but add benefit for desktop users
  2. adding optional service supervision & monitoring will not hurt you, - " - for server use-case
  3. It will not fundamentally change FreeBSD
I do not want to replace FreeBSD. I propose to enhance it.
I would even consider myself a "non-nerd" user. I am not an OS developer. Heck, I am a slow tech learner which is probably why I think Linux is defective. I just use FreeBSD to provide a platform to develop software and teach on. Windows and Linux have so much breakage daily that they no longer provide a solution to my needs. How is that user-friendly?
Click to expand...
Your implication that I want FreeBSD to become like Windows or Linux is not right.
The sad truth is that these haven't seen much uptake because they don't offer more than systemd or existing standard init systems. There are a number of arguments why they are not popular, just do a search for the "systemd vs openrc" threads. In terms of FreeBSD they do not offer enough to justify the breakage they would cause.
Click to expand...
OK now I will suspend with automagic hibernation after 15 minutes, then reboot, and measure the times to startup the base system, and until the X11 login screen. I have to set up another boot env anyways, I'll try to replace it's init with runit(8). Anyone who's curious and open-minded to check such tool unbiased, can decide to do so as well.
kpedersen said:
I expect zoujiaqing has a lot of work ahead of him :)
Click to expand...
That's why I came up with my suggestion. Iff - and only if - such replacement does not break reliable & proven BSD policies, this work would be done once in the base for those who want it, and not hurt you & the others.
 
I like runit. But paralising init system will not speed up boot times, when hardware detection, e.g. USB, is the issue. Or an ntpdate which takes a few seconds.
I ask myself can you combine runit/daemontools/rc ?
 
mjollnir said:
They can decide for themselves... Ordinary people want a desktop computer system to write e-mail, chat, browse the internet, etc.pp. Note the act of writing.
Click to expand...

If for these "ordinary people" FreeBSD is not good for that purposes, if they are unhappy with FreeBSD, then they should use Windows or Ubuntu that are made for them and are perfect
for their act of writing.

What about ordinary people that want stay conscious of what the system does and eventually
alter it? That want to write and run small programs, for example for technical, scientific,
commercial or other purposes? People that see in a classical computer a computer and not an
application box?

Why are you zealously defining "ordinary people", insistingly speaking in their name, and ignoring
the existence and needs of other "ordinary people"?

As I wrote in other thread, OpenBSD is more coherent. Its rc is in my opinion much
simpler as also the whole system. If you were in the OpenBSD mailing list, you would get
the answer you deserve.
 
hruodr said:
If for these "ordinary people" FreeBSD is not good for that purposes, if they are unhappy with FreeBSD, then they should use Windows or Ubuntu that are made for them and are perfect
for their act of writing.
Click to expand...
I already answered this above. Because if FreeBSD was more widely deployed & known as a stable, reliable (in our opinion: superior) alternative to Windows & Linux, this would result in more developers & maintainers after a few years.
What about ordinary people that want stay conscious of what the system does and eventually
alter it? That want to write and run small programs, for example for technical, scientific,
commercial or other purposes? People that see in a classical computer a computer and not an
application box?
Click to expand...
I do care as well. I am one of these.
Why are you zealously defining "ordinary people", insistingly speaking in their name, and ignoring
the existence and needs of other "ordinary people"?
Click to expand...
Why are you suspecting (accusing) me that I want to "take away my OS"? Please let's stay unemotional, factually. Provide me an argument why parallel service startup (off by default) would take away any of the beloved simplicity & rubustness of FreeBSD.
 
mjollnir said:
rovide me an argument why parallel service startup (off by default) would take away any of the beloved simplicity & rubustness of FreeBSD.
Click to expand...

Many people told it: non deterministic behaviour, chance of race conditions, complexity, no need
to make startup faster and not at this price, system startup is a critical moment and should be
kept as simple as possible.

mjollnir said:
Because if FreeBSD was more widely deployed & known as a stable, reliable (in our opinion: superior) alternative to Windows & Linux, this would result in more developers & maintainers after a few years.
Click to expand...

That is a very primitive, one dimensional way of thinking: more deployed and known, then
superior, then more mantainers and developers. The more, the better. For people that think
that way there are more intelligent, and hence better operating systems that think for
them: for example Windows and Ubuntu.
 
mjollnir said:
Why are you suspecting (accusing) me that I want to "take away my OS"?
Click to expand...

I guess it is more the worry of feature creep. As in once we have parallel jobs (off by default), what happens if someone decides to make it on by default and breaks some niche legacy software you need to run?

But you are correct, improving it doesn't need to break anything. I.e if it was first implemented as a drop in replacement with many compatibility modes so that none of the existing subsystems or ports break and then slowly update them one by one to the new (i.e parallel) functionality until none of the legacy stuff remains.

This is the only responsible way to replace large systems. Something that often gets overlooked in Linux in the name of "blind" progress. But this blind progress is so annoyingly damaging when it does occur. Us FreeBSD guys like to stay out of all that when possible.
 
I ran away from Linux & Solaris to FreeBSD for exactly these reasons.
hruodr said:
Many people told it: non deterministic behaviour, chance of race conditions, complexity, no need
to make startup faster and not at this price, system startup is a critical moment and should be
kept as simple as possible.
Click to expand...
  • non-deterministic behaviour is no problem if the dependencies are correct
  • race conditions do appear in the current system as well; parallel/serial execution can be switched easily
  • dependencies are already encoded in the dependency hacks, thus complexity will very likely not expand too much
  • many desktop users are asking for parallel service startup
  • the two existing systems I mentioned are reported to be fairly KISS
That is a very primitive, one dimensional way of thinking: more deployed and known, then
superior, then more mantainers and developers. The more, the better. For people that think
that way there are more intelligent, and hence better operating systems that think for
them: for example Windows and Ubuntu.
Click to expand...
You are going emotional?
 
Why to expand complexity if it could also be reduced? And for you is non-deterministic behaviour
not a problem, for others is even that execution time cannot be determined a big problem. But
let us take technical considerations away for a while:

mjollnir said:
many desktop users are asking for parallel service startup
Click to expand...

How many desktop users you mean? From where you know them? Do they asked you for that?
And what services do these desktop users start at boot time? Or they asked for a feature without
knowing for what they need the feature?
 
hruodr said:
How many desktop users you mean? From where you know them? Do they asked you for that?
And what services do these desktop users start at boot time? Or they asked for a feature without
knowing for what they need the feature?
Click to expand...
Naturally, the number of whom I personally know is much lower of those I pretend to speak for. Friends & family. Collegues. Myself. People I do not really know well, but I see them using a Linux desktop. Observations on various forums. Yes, they do ask when I suggest FreeBSD as a superior alternative to Linux to them; either they want to move away from Windows or Mac to Linux, or they are not satisfied with their current Linux distro. Then two issues come up over and over again:
  • suspend & resume: this has greatly improved on FreeBSD and modern hardware became more conformal to standards
  • parallel service startup, i.e. time from power-on to GUI login screen; even with a SSD, FreeBSD is factors slower than it could be; Linux is much faster
Most of them start XfCE, Gnome, or KDE desktop environment with no bells & whistles.
 
You must log in or register to reply here.
Back
Top