If a service doesn't quite fit systemd, work on systemd until it fits, don't rewrite it!
Part of the appeal of the systemd is that it minimises the idiosyncrasies of all the different tools. You don't have to rewrite the services, but fitting them to systemd is reasonable. For example, udev was not rewritten, just integrated.
originals were made by experts in that field
Whatever made you say that? There are no mandatory security audit, secure coding practices, mandatory code reviews, qualification certificates or anything like that in free software. Some of it is even insecure by design (I'm looking at you, X.org). systemd will get rid of
So because all software has bugs we should just say: "Meh, who cares if systemd is bloated, has monumental feature creep, an author with a serious ego problem who doesn't like fixing them and it replaced an init which, while a bit crusty did what it said on the tin, because, well, systemd." Right?
Wrong. Systemd, is a poor idea, poorly implemented in a project thats poorly managed. Quite why redhat are in thrall to the arrogant little twerp in charge is anyones guess.
by Anonymous Coward writes:
on Monday July 03, 2017 @05:58AM (#54733511)
Systemd, is a poor idea, poorly implemented in a project thats poorly managed.
Yup.
Quite why redhat are in thrall to the arrogant little twerp in charge is anyones guess.
I don't know how he got a foot in the door, so all I have is my long-distance high-level. The short and sweet is "politics", and he's red hat's useful idiot in their war with oracle.
red hat tried to tighten up their business model by splitting red hat-the-distro into fedora and RHEL. The centos guys forced red hat to release RHEL anyway because GPL. oracle then copied the whole thing verbatim, minus the branding, and called it "unbreakable linux". This allowed them to piggyback on red hat's hard work including their knowledge base. Which is why red hat put that thing behind a paywall. So oracle is doing their level best to eat red hat's lunch and using red hat's own work to do it.
So, what can red hat still do? Well, they can become more proprietary, which is where systemd comes in. No, it's not closed source, but it sure is wilfully, deliberately, and very gratuitously, incompatible with everyone else. So if a big enough party like red hat pushes it into the market, then the rest (*cough* debian *cough*) feels they have to follow suit. That means a big boon in control for red hat. And that's what it ultimately is about.
Of course, you have to have some sort of story of benefit for people to go along with the gambit. But for many the sort of fine-grained control you lose by going systemd wasn't that important anyway, for the good and simple reason that throwing more virtualisation and more hardware at it seems faster and so more cost effective. It's the same sort of reason why virtualisation and massive hardware is so popular in the windows world. The big gain for red hat sure isn't technical, it's political. You can see this in the fact that many rebuttals to technical complaints about systemd are entirely non-technical. Any time a technical argument ends up full of bullshit and personal attacks it really isn't about the technical merits any more.
I doubt poettering himself understands his role in this, seeing his grasp of architecture, code quality, and so on, so that makes him a "useful idiot". He sure does have the personality to pull it off, though. Incompetent arrogance does go a long way, with the right backing.
And, of course, once they get off the ground, bad ideas are hard to kill.
Quotes, with minor modifications to make comments more readable, and some text in bold:
"Systemd is a poor idea, poorly implemented, in a project that's poorly managed."
"... sure is wilfully, deliberately, and very gratuitously, incompatible with everyone else.... That means a big boon in control for Red Hat.
"I doubt Poettering himself understands his role in this, seeing his grasp of architecture, code quality, and so on, so that makes him a 'useful idiot'. He sure does have the personality to pull it off, though. Incompetent arrogance does go a long way, with the right backing."
Systemd, and the poor way it has been presented, had been damaging to Red Hat's reputation ("Dead Hat"). Why did Red Hat management allow that? Is it because Red Hat makes money providing consulting services, and wants Linux configuration to be difficult so that the company will make more money?
Why did Red Hat management allow that? Is it because Red Hat makes money providing consulting services, and wants Linux configuration to be difficult so that the company will make more money?
Original emails concerning "why systend" have been published. Years ago. All have read them I trust. The given reason was to force **brand** on RedHat/Gnome products.
Systemd, is a poor idea, poorly implemented in a project thats poorly managed.
If systemd is bad, why the distros featuring it are thriving, while the non-systemd ones are small projects with uncertain future [Nota bene: I'm not a systemd advocate?
Inertia. The specifically non-systemd ones are either brand new and are in the early stages of adoption/development (Devuan) or are old, but haven't had a large following for years (Slackware). [Slackware will never die, by the way. It's one of the immortals.]
Let's look at it another way: This may be because I'm not in the corporate world, but I haven't met somebody who has deliberately installed anything from RedHat (or had anything good to say about RedHat) during this decade and systemd development and i
Most shops I know that are paying for RHEL support contracts, not for the software. That gives them somebody to call if issues come up or there's a critical bug to fix, as well as access to the KB's etc.
I don't know that SystemD changes this much. If anything, it's hurting RedHat's reputation.
The problem with systemd (Score:5, Insightful)
That's a problem with Systemd. It's a pretty decent idea with a sub-par execution and a crappy way of dealing with an inherent problem.
Idea: centralized place to optimize startup, management and interconnectivity of all kinds of services.
Problem: some services in their standard form don't quite fit that model.
Solution: let's rewrite them and include as parts of systemd.
The crap part: while the originals were made by experts in that field, the replacements are made by a group of wannabe experts on everything
Re: (Score:3, Informative)
If a service doesn't quite fit systemd, work on systemd until it fits, don't rewrite it!
Part of the appeal of the systemd is that it minimises the idiosyncrasies of all the different tools. You don't have to rewrite the services, but fitting them to systemd is reasonable. For example, udev was not rewritten, just integrated.
originals were made by experts in that field
Whatever made you say that? There are no mandatory security audit, secure coding practices, mandatory code reviews, qualification certificates or anything like that in free software. Some of it is even insecure by design (I'm looking at you, X.org). systemd will get rid of
Re: (Score:5, Insightful)
"Some of it is even insecure by design "
So because all software has bugs we should just say: "Meh, who cares if systemd is bloated, has monumental feature creep, an author with a serious ego problem who doesn't like fixing them and it replaced an init which, while a bit crusty did what it said on the tin, because, well, systemd." Right?
Wrong. Systemd, is a poor idea, poorly implemented in a project thats poorly managed. Quite why redhat are in thrall to the arrogant little twerp in charge is anyones guess.
Re:The problem with systemd (Score:5, Interesting)
Systemd, is a poor idea, poorly implemented in a project thats poorly managed.
Yup.
Quite why redhat are in thrall to the arrogant little twerp in charge is anyones guess.
I don't know how he got a foot in the door, so all I have is my long-distance high-level. The short and sweet is "politics", and he's red hat's useful idiot in their war with oracle.
red hat tried to tighten up their business model by splitting red hat-the-distro into fedora and RHEL. The centos guys forced red hat to release RHEL anyway because GPL. oracle then copied the whole thing verbatim, minus the branding, and called it "unbreakable linux". This allowed them to piggyback on red hat's hard work including their knowledge base. Which is why red hat put that thing behind a paywall. So oracle is doing their level best to eat red hat's lunch and using red hat's own work to do it.
So, what can red hat still do? Well, they can become more proprietary, which is where systemd comes in. No, it's not closed source, but it sure is wilfully, deliberately, and very gratuitously, incompatible with everyone else. So if a big enough party like red hat pushes it into the market, then the rest (*cough* debian *cough*) feels they have to follow suit. That means a big boon in control for red hat. And that's what it ultimately is about.
Of course, you have to have some sort of story of benefit for people to go along with the gambit. But for many the sort of fine-grained control you lose by going systemd wasn't that important anyway, for the good and simple reason that throwing more virtualisation and more hardware at it seems faster and so more cost effective. It's the same sort of reason why virtualisation and massive hardware is so popular in the windows world. The big gain for red hat sure isn't technical, it's political. You can see this in the fact that many rebuttals to technical complaints about systemd are entirely non-technical. Any time a technical argument ends up full of bullshit and personal attacks it really isn't about the technical merits any more.
I doubt poettering himself understands his role in this, seeing his grasp of architecture, code quality, and so on, so that makes him a "useful idiot". He sure does have the personality to pull it off, though. Incompetent arrogance does go a long way, with the right backing.
And, of course, once they get off the ground, bad ideas are hard to kill.
Re: (Score:2)
Can't disagree with any of that.
Why does Red Hat allow damage to its reputation? (Score:5, Insightful)
"Systemd is a poor idea, poorly implemented, in a project that's poorly managed."
"... sure is wilfully, deliberately, and very gratuitously, incompatible with everyone else.
"I doubt Poettering himself understands his role in this, seeing his grasp of architecture, code quality, and so on, so that makes him a 'useful idiot'. He sure does have the personality to pull it off, though. Incompetent arrogance does go a long way, with the right backing."
Systemd, and the poor way it has been presented, had been damaging to Red Hat's reputation ("Dead Hat"). Why did Red Hat management allow that? Is it because Red Hat makes money providing consulting services, and wants Linux configuration to be difficult so that the company will make more money?
Re: (Score:1)
Why did Red Hat management allow that? Is it because Red Hat makes money providing consulting services, and wants Linux configuration to be difficult so that the company will make more money?
By George, I think he's got it!
Re: (Score:1)
Original emails concerning "why systend" have been published. Years ago. All have read them I trust. The given reason was to force **brand** on RedHat/Gnome products.
Re: (Score:2)
Systemd, is a poor idea, poorly implemented in a project thats poorly managed.
If systemd is bad, why the distros featuring it are thriving, while the non-systemd ones are small projects with uncertain future [Nota bene: I'm not a systemd advocate?
Inertia. The specifically non-systemd ones are either brand new and are in the early stages of adoption/development (Devuan) or are old, but haven't had a large following for years (Slackware). [Slackware will never die, by the way. It's one of the immortals.]
Let's look at it another way: This may be because I'm not in the corporate world, but I haven't met somebody who has deliberately installed anything from RedHat (or had anything good to say about RedHat) during this decade and systemd development and i
Re: (Score:1)
Most shops I know that are paying for RHEL support contracts, not for the software. That gives them somebody to call if issues come up or there's a critical bug to fix, as well as access to the KB's etc.
I don't know that SystemD changes this much. If anything, it's hurting RedHat's reputation.