Systemd Gains IP Forwarding, IP Masquerading & Basic Firewall Controls

[pal666]

Should we start to call the OS Systemd/Linux instead of GNU/Linux?

gnu is multiplatform, while systemd is linux

[ultimA]

Systemd is designed, and will replace most of the Linux base system - especially when dealing with light-weight VM's and containers.
As an embedded system developer, I for one welcome my new systemd overlord.

Don't like it? Think systemd is hurting your personal freedoms? I'd suggest you join Devuan team.
Maybe, just maybe, if sufficient people will stop wasting time trashing systemd in Slashdot/Phoronix/IWN and get hacking instead, Devuan will actually prove to be an viable systemd alternative.

- Gilboa

Two things:

First, I've been using systemd (part of multiple archlinux setups) myself, and it's been working fine for me. That doesn't mean though I agree with its design principles, since as a long-time developer, my experience has been that some design-decisions of systemd will backfire at some point in the future. This is also the point about systemd many others dislike (and yes there are also some angry sysadmins who now need to adapt, but I feel this is a smaller part of the critisism).

Second, I humbly reject all posts that basically say "if you don't like it, go build your own distribution/init/whatever". Because as a user, especially power user, I have every right to critisize it and let my like OR dislike be known, even if I am not working on that project or any of its alternatives. Everybody who tells me that, should first build their own car/smartphone or shoot their own movie before being allowed to critisize another one.

[Krejzi]

It's just a front-end to iptables, so what's the big deal? It's not like they reimplemented the entire stack or integrated iptables into systemd.

[prodigy_]

Should we start to call the OS Systemd/Linux instead of GNU/Linux?

No, simply Red Hat OS.

I can remember, many months ago somebody mentioned systemd will at one point implement firewall control. Except, at that time it was meant as a joke.

Kerneld is on the way.

[Brane215]

While I like systemd as a concept, implementation kinda sucks.

Especially networking part. What's wrong with doing full copy af already existing DHCP utils into systemd, tweaked to be executed as a loadable module instead separate executable ?

DHCP and other parts are operating in "limp mode" - good enough for 30 second demonstration, but totally braindamaged for real use.

same with networkd. You don't have any influence over its work other than link and network files. So this means that if network link gets botched, you don't have any means to restart it through networkd.

Or to change network configuration etc.

[stalkerg]

If this replace iptables - NOT NEED.
If this ui to iptables - YOU BROKE MY RULES!

Systemd please https://en.wikipedia.org/wiki/Ouroboros .

[pingufunkybeat]

I can remember, many months ago somebody mentioned systemd will at one point implement firewall control. Except, at that time it was meant as a joke.

Systemd is not ready until it implements the complete OpenGL spec.

This time next year, Mesa will be pulled into the systemd tree, and you'll be asked to fork it if you don't like it

[Gusar]

It's just a front-end to iptables, so what's the big deal?

It involves systemd. So it of course must be blown out of proportion.

same with networkd. You don't have any influence over its work other than link and network files. So this means that if network link gets botched, you don't have any means to restart it through networkd.

That's just the current state of things. It was said from the beginning that there will be an interface to control networkd at some point.


@ultimA: The problem is, most criticism is of crappy quality. And that a lot of it is just repeating the same old drivel that has not practical value, it's just noise.

[gilboa]

Two things:

First, I've been using systemd (part of multiple archlinux setups) myself, and it's been working fine for me. That doesn't mean though I agree with its design principles, since as a long-time developer, my experience has been that some design-decisions of systemd will backfire at some point in the future. This is also the point about systemd many others dislike (and yes there are also some angry sysadmins who now need to adapt, but I feel this is a smaller part of the critisism).

Oh, OK.
I never claimed systemd to be remotely close to perfect.
I do claim to developing and fielding complex multi-service appliances is now far easier, than it used to, back in the (not-so) good old SysV days.

Second, I humbly reject all posts that basically say "if you don't like it, go build your own distribution/init/whatever". Because as a user, especially power user, I have every right to critisize it and let my like OR dislike be known, even if I am not working on that project or any of its alternatives. Everybody who tells me that, should first build their own car/smartphone or shoot their own movie before being allowed to critisize another one.

I'm not sure what's your point.
If you enjoy drawing criticism, and it somehow makes you feel better, please don't let me stop you.
Though, given the fact that Linux (and the rest of GNU) was founded by people with an itch to scratch (and based on these principles), don't be shocked if your criticism is largely ignored.

... On the other hand, if you want to *do* something about that systemd scratch that's itching you (as opposed to generating white noise), join the Devuan team.

- Gilboa

[stalkerg]

Systemd is not ready until it implements the complete OpenGL spec.

This time next year, Mesa will be pulled into the systemd tree, and you'll be asked to fork it if you don't like it

"Mice were crying, injected, but continued to chew cactus."