Another thing I am noticing is the NSA is being used as a Red Herring. Everyone focuses on the "Big Bad NSA"... Meanwhile everyone forgot about the real blackhats that run things underground as well as the big Corps that can buy the NSA and all its agents.

Coming from this point of view you really can't trust any code you did not write your self. (I think that's Ken Thompson's quote)

If only I could trust code I write myself

I shared this video link, because I've seen cheap emotional/psychological manipulation used inside this forum, too. So while PHK humorously pointed at the NSA, there a people out there who really use these tactics to push their agenda against Linux.

Of course, my trust in FOSS and the "community" is completely gone already and this has consequences. So people talking like "you will accept systemd, because it's mandatory" followed up later by "now you will accept systemd-cloud", and then "GNOME4 now requires systemd-sshd". Somewhere I have to draw the line, and mine is right here, just before PID 1 gets replaced by PulseAudio++.

1 members found this post helpful.

AFAIK the biggest reason corps r pushing FOSS is to get "Free Labor and lower costs". It never offered me a sense of security from the beginning. Do you even trust the slackbuilds packages? I don't. Same with all the other precompiled packages in the bigger distros as well as slackware cd itself. Open or closed source doesn't really make too much of a big difference when you use precompiled software. Besides most source code is so long it take hours of time just to check for bugs or trickery in it. Imagine having to go through all the source code for Linux? Millions of lines of code... Good luck! Often the worst holes are in the least expected places too.

So the situation is like an Onion. There are layers and layers of insecurity in the game.

Na I didn't mean it like that. i meant because every other distro has accepted it including the final big elephant Umbuntu. So that being said I honestly don't think linux devs are going to write to earlier versions of init.

Where to? You can't really escape it in the internet world. Where you gonna go to?

I patiently await systemd-shell, systemd-kernel, and systemd-cc. Then we'll have a complete LennartOS.

3 members found this post helpful.

After reading through some of the various threads on systemd by Pat, it seems like he's rather pragmatic about the whole thing. He doesn't relish the idea of it, but he will integrate it if he has to.

Pro-systemd people always post this link: http://0pointer.de/blog/projects/why.html. Here's my thoughts:

• Shell free bootup. One of the nice things about shell scripts is they're very transparent and easily hackable. You can say service files are the same, but because systemd is attempting to do additional things like cron, syslog, and bring up the network, you do not have the same level of transparency into the system. If systemd doesn't configure my network correctly before trying to bring up nfs mounts, how are you going to fix it? Filing a bug report will take time, and if you are inclined to patch it your patch may not be accepted upstream.
• Claiming that init can't do half the things systemd does. Want to know why? It doesn't need to! Most of the items listed can already be done with the utilities systemd is trying to replace. LUKS handling? Not a problem in sysvinit either!
• Disabling services w/o editing files. Is it really that hard to delete a symbolic link or run a GUI like system-config-services?
• Reliable termination of user sessions before shutdown. So kill doesn't work anymore?
• "Easily writable, extensible and parseable service files, suitable for manipulation with enterprise management tools." So there aren't any GUIs whatsoever to handle init scripts? I think RedHat would like to have a word about that since they created this: https://access.redhat.com/site/docum...nfiguring.html
• Thank goodness for this: "Specialized professional consulting and engineering services available." Do we need this for sysvinit? Didn't think so.

Now let's look at some arguments external to Lennart's site:

• http://ewontfix.com/14/.
• Support. If you find a bug with systemd, will it get closed out for being "working as designed?" Take a look at https://bugzilla.redhat.com/buglist....ponent=systemd and see some of the responses. Because of the simplicity of init, you shouldn't find too many bugs even if you get no support.
• Repairability. If I have an init script that causes a kernel panic, I can boot up a Linux USB drive, mount the filesystem, and remove the link to the file in /etc/rc?.d so it doesn't run. If I have a service file in systemd that causes a kernel panic, I must have a running Linux system first to run systemctl on. If I delete the service file outright, would systemd still try to start it and hang up? How do you deal with this?
• Binary log files. This reduces interoperability with any other programs that might want to interact with systemd, as they are reliant on journalctl. I can see this being a problem for an enterprise logging server trying to pull logs from mulitple servers in near-realtime, like LEM.

Now some pros to systemd:

• Service dependencies. I liked this in Solaris as well, although it is hard to write services for SMF.
• Parallel boot. This can be a good thing, but parallelized boot can cause strange race conditions that would be hard to diagnose.
• CGroups support. More flexibility is good.

4 members found this post helpful.

What is that what I said?

Then you gonna have a bunch of people that believe you without reading what I originally wrote and start flaming. OK I am out...

I thought you were stating here that anti-systemd people were zealots. I apologize if that was not your intention.

No problem. But yes thats correct. I didn't mean "anti-systemd" people. I meant FUD spreaders with comments that don't make any sense. I don't want to point people out but you also probably know. I have nothing wrong with tech arguments but some of the stuff I have to read through is just ...ughhh.

Infact I am not Pro SystemD. All inits have its ups and downs. I actually prefer sys v over systemd for my desktop use. My message is for those that just spread FUD for the sake of FUD.

SystemD is Powerful... but it makes things more complicated in certain situations. But I am not AntiSystemD either. I'm neutral about it really.

Took out that sentence, sorry about that. This thread wasn't supposed to be about pro/against systemd anyway, but it derailed pages and pages ago.

Threads like this are going to go tangential because it is a foggy subject at best and one loaded with both "sacred cows" and real concerns. I have asked in many forums for a simple answer to what systemd's benefits are besides faster boot times. I have yet to get a solid answer. I, and I assume others, might not be listed in the "anti camp" if we could just see that it is not actually a solution seeking a problem, and frankly, boot time is not sufficient for a decent "return on investment".

After spending easily a hundred+ hours pouring over blogs and websites, I think part of the problem of trying to talk intelligently about systemd originates with the 2 main developers, Poettering and Sievers. They deny obvious bugs and abandon maintenance, blame everything on everyone else, and blatantly state outright "you're either with us or against us" and generally behave in some shady, irresponsible and arrogant ways. Admittedly I have some problems with accepting "integrity" like that messing with PID 1 on my box, but that is somewhat beside the point. That sort of confrontational attitude and wanton dismissal of responsibility makes it all but impossible to talk about it intelligently until after it is fait accompli!

I do expect Debian will work very hard to make it viable and let's face it, RHEL 7 will be under heavy scrutiny, so apparently they have some sort of plan that doesn't involve them losing even more market share from irate clients at the doors with torches and pitchforks.

Until recently I was vaguely against Wayland. I watched a single video on YouTube of a talk by one of the devs, Daniel Stone, and he made some really good points. He clearly stated problems in X, many of which I have experienced, and what they have come up with as solutions. He even goes on to talk about problems with X for which they have yet to find solutions "because it's really hard". This man is not blaming anyone else, displays responsibility, is not arrogant, dismissive or confrontational and I want to see what they produce. I'm looking forward to it.

So it would seem that I'm not a zealot or conspiracy theorist. I can be convinced to overcome my prejudices. Why can I not find one single webpage, person, symposium, whatever that can lay it out like the above? In the absence of honesty, responsibility, openness and integrity, I am left only to my imagination because there can be no trusting, no confidence, absent those characteristics.

Given Patrick and Contributors track record, I am confidant that if Rome is going to burn, we will be watching safely from the mountains. We won't "do as in Rome" until there is demonstrable reason to.... They will have to be convinced and barbarians wearing Senate robes ain't gonna do it.

13 members found this post helpful.

^^^

Well said!

Exactly that is the reason why RH is now controlling almost the complete ecosystem. I don't want to be part of that.

The TINA argument. I hear that repeated again and again. It's just plain wrong. There is always the option to switch sides.

Didn't Windows 7 and 8 both migrate somewhat away from the reboot to upgrade feature? I think so because you can perform a lot of system and driver updates sometimes without needing to reboot, as it just restarts the service handler rather than the whole underlying system.

It's a shame Windows is moving more towards what UNIX does by design while Linux is moving towards what Windows used to be like.

Sadly Enorbet, Wayland was developed to actually fix issues that were known to be broken and had been broke for quite some time with no answers.

Systemd was designed to replace an already working subsystem and collection of tools that were not broken and not in need of heavy repair work on any level.

And yes, you hit the nail square on the head about how Lennart and Kay treat FOSS projects they are involved with. The arrogance they assert was even mentioned by Ubuntu regarding PulseAudio's early inception. It wasn't broken because of Ubuntu's implementation, it was broken by design.

2 members found this post helpful.

So what you gonna do write a new OS?
You guys are so focused on RH that you forget Oracle, Microsoft, Apple, Google, yada yada yada. Or like I said earlier what makes you so secure about the Linux Kernel its self? Its only a couple million lines of C code... ?

...PS - RH has more fingers on the Kernel then any other distro I know.