Systemd's Plan For Stateless Systems, Factory Resets

[paradis]

...
Are these things I should not care about?

Linux systems that just work is a nightmare for many. systemd is perhaps the biggest threat to linux job security in many years,
this is the #1 reason for all the angry and desperate trolls.

[psychoticmeow]

Systemd was an inside job MAAAAAN.

Idiot.

[r_a_trip]

Linux systems that just work is a nightmare for many. systemd is perhaps the biggest threat to linux job security in many years,
this is the #1 reason for all the angry and desperate trolls.

I'd say that Linux systems that "just work" are an opportunity to generate more jobs. If Linux becomes easier to administer and more reliable, it stands to reason more companies will deploy Linux solutions, which will need more Linux administrators. What won't be needed anymore are primadonna "old school" Linux admins, who have deployed their own homegrown solutions, which can only be administered by (other) primadonna "old school" Linux admins. Overall the pie gets bigger, but the slices might get a bit smaller.

[jonnor]

You should care about whatever is relevant for you. (For me none of those things is even remotely significant, sorry.)

But everything comes for a price. The price of systemd is... Linux as we know it. All the things we were promised, such as "freedom of choice" and "Linux is not a corporation" are flying completely out of the window right now - do you feel it?

I do not feel it. And feelings are not good arguments in a discussion.
With systemd you have the 4 freedoms as defined by the Free Software Foundation. What other freedoms where you promised and by whom, and how is systemd a threat to these?

[jonnor]

Linux systems that just work is a nightmare for many. systemd is perhaps the biggest threat to linux job security in many years, this is the #1 reason for all the angry and desperate trolls.

It is perhaps perceived so by some people. But making Linux a significantly better technical solution to embedded, mobile and server+cloud usecases will make Linux adoption go up and the job market for Linux experts likewise.
We may see a slight shift in required skillsets though, but these are not driven primarily by systemd - but the same external factors that systemd are responding to. In particular, "devops" and other people with competence in continuous integration and deployment will be favored over traditional "sysadmins" who handmanage or script maintenance of VMs or physical servers. As the core OS is more and more a commodity, there might also be a shift towards Linux "application developers" compared to "system developers", particularly in embedded devices and backend work for web.
Third-party dev for desktop and mobile looks more unlikely, we still don't have a widely adopted GNU/Linux platform there.
But there is the hope that with appropriate technical solutions, we can grow third-party ecosystems for (self-hosted) server+cloud and for embedded devices (home automation, automotive), which would be a significant improvement in the freedom of users compared to now.

[FLHerne]

So may I suggest that we stop calling it "Linux" and call it Red Hat OS instead? Simply to avoid confusion.

You can suggest it, but I don't think you'll get many takers.

There is no RedHat conspiracy to force people to use systemd. I voluntarily choose systemd over Upstart/OpenRC (sysvinit isn't even a serious contender), because it's simply better and makes things simpler and more reliable.

Apparently most distro maintainers agree with me. Systemd hasn't been carefully hacked into their repositories by the RedHat conspiracy, it's been chosen by their maintainers because it makes their work easier and the distro better. Applications don't magically get sucked into having systemd dependencies, their authors have just found that relying on systemd rather than ugly hacks makes them more reliable.

One more time: you really, really should stop talking on behalf of everyone [who uses Linux]

FTFY, and nope.
Since you don't actually seem to use Linux (your repeated "systemd does nothing for me, despite simplifying things that every Linux user does regularly" claims), everyone else choosing to use systemd won't affect you at all.

And if you do occasionally want to post nonsense on the interwebs, there's always OpenRC. Sure, it doesn't do all the things systemd does, but as you keep pointing out you don't use those things anyway...

[prodigy_]

Since you don't actually seem to use Linux

I still use Linux at the moment but I'm halfway to BSD so to speak. One must always hedge against risks and risks associated with over-relying on Linux are currently at all time high.

People say that I'm a hater or a troll. The truth is that I'm neither. Life has taught me not to break things that work - at least not without a very good reason. So I'll continue to observe this systemd revolution. It's going to be fun. Especially from a safe distance.

[interested]

Very interesting.

Will these new sandboxing features make security systems such as SELinux or RBAC obsolete ?

No, they will work on top of (or together with) RBAC's and MAC's like SELinux, they have different aims and work in different ways.

Remember though, that the app sandboxing is still in the future with different ideas of implementation, and at the moment, mostly a Gnome project. KDE or others may come up with competing sandboxing solutions. A cross distro sandboxing SDK with a stable API may not materialise any-time soon.

But it will be awesome if it does; there are +80 million Smart TV's sold every year and they probably all run Linux. But each vendor has its own eco system for apps. While HTML5 will solve some problems, Netflix and many other would really like if one app would work across all 80 million TV's.

I think the future Linux distros gradually will increase sandboxing of both processes/services and applications making a much tougher defence in depth against attacks. Even partial sandboxing, like systemd's "NoNewPrivileges=" and "ReadOnlySystem=" will dramatically increase service security with very little user impact. With Wayland there will be even further process isolation, eg. a malware weather app running in the desktop panel, won't be able to scrape passwords and data from other parts of the screen etc.

[TheBlackCat]

Apparently most distro maintainers agree with me.

And desktop environment maintainers.

[TheBlackCat]

Life has taught me not to break things that work - at least not without a very good reason.

Except that pretty much everyone in a position to make decisions in the Linux ecosystem, including pretty much all distro and desktop environment maintainers, think there was a very good reason.

And you still haven't explained why you think this is the case. How is it that you are right while pretty much anyone who has earned a position of trust in the Linux ecosystem is wrong?