Debian To Switch To Systemd Or Upstart

[dh04000]

For a nice wiki-style debate with pros and cons for the different options, see:

https://wiki.debian.org/Debate/initsystem

far less polarized than the phoronix debate

Personally I think openrc would be a good match, but we will see what they decide

My favorite quote so far in the Debian debate;

Le vendredi 10 ao?t 2012 ? 17:04 +0900, [email protected] a ?crit :
> Debian is about the freedom to choose.

No, it is not.

Lol, wat?

[ArneBab]

The ideas behind systemd are pretty strong:

I don?t like about systemd, that it seems to make it harder to customize the init process. It means that the systemd devels get a position of power over the distros: Either there is someone who can integrate some functionality in systemd, or a given feature cannot be provided (I blogged a slightly longer version of this: http://draketo.de/light/english/top-5-systemd-troubles ).

With systemd, there is one kernel and one way to start the system - and multiple distros.

[dh04000]

I don?t like about systemd, that it seems to make it harder to customize the init process. It means that the systemd devels get a position of power over the distros: Either there is someone who can integrate some functionality in systemd, or a given feature cannot be provided (I blogged a slightly longer version of this: http://draketo.de/light/english/top-5-systemd-troubles ).

With systemd, there is one kernel and one way to start the system - and multiple distros.

Doesn't seem fair to lock the kernel down to just ONE way of doing something. Should we lock the rest of it down to ONE window manager and environment? NO! We would just have the linux version of Windows then.

[erendorn]

I don?t like about systemd, that it seems to make it harder to customize the init process. It means that the systemd devels get a position of power over the distros: Either there is someone who can integrate some functionality in systemd, or a given feature cannot be provided (I blogged a slightly longer version of this: http://draketo.de/light/english/top-5-systemd-troubles ).

With systemd, there is one kernel and one way to start the system - and multiple distros.

That looks quite similar to the situation of the kernel itself, though.

[ArneBab]

That looks quite similar to the situation of the kernel itself, though.

Yes - but this situation is changing. Even though Linux has the huge wave of success (with android) it is becoming ever easier to test alternate kernels with the distribution you know.

[gens]

Yes - but this situation is changing. Even though Linux has the huge wave of success (with android) it is becoming ever easier to test alternate kernels with the distribution you know.

oh, i thought he meant "systemd is doing the kernels job"

[Ibidem]

That looks quite similar to the situation of the kernel itself, though.

I dare you to:
1: get the latest kernel source
2: cd into the root of the source tree
3: type

make config

and answer each question without just accepting the default
4: build, install, and boot the resulting kernel.

Be sure to allow a day or two for it.

[chithanh]

That is a pointless exercise, and has nothing to do with the init system.
You start with "make defconfig" or "make localyesconfig", and then enable the things you additionally need in menuconfig.

[Awesomeness]

Lol, wat?

Debian was never about choice. If that was the case, ffmpeg and cdrecord were also in the repos, not only libav.org and cdrkit.

[oleid]

I don?t like about systemd, that it seems to make it harder to customize the init process.

How does it do that?

When I was forced to install systemd on my gentoo due to the installation of gnome 3.8, I didn't like it. But after reading about the usage, I see no reason to dislike it. Works like a charme.

Actually, I didn't find out how to setup a bridged network - the former gentoo way seemed not to be supported, so I just hacked all the commands I need to set up the very network infrastructure into a shell script and ordered systemd to launch that file in order to set up the network... Very easy.

It means that the systemd devels get a position of power over the distros: Either there is someone who can integrate some functionality in systemd, or a given feature cannot be provided

This fear seems quite artificial. Basically all distributions have the same needs: launching daemons and logging in users. What is different is - in most cases the path of the binary of a daemon and the parameters, which are passed. And this is controlled via simple text files.

(I blogged a slightly longer version of this: http://draketo.de/light/english/top-5-systemd-troubles ).

I don't know how hard the dependencies on the kernel version are, but typically, I am using a recent version anyway - whatever Gentoo declares stable.
Non-rolling distributions don't upgrade the init system during the life cycle of a release, so I don't see a problem here, too.

Customisation via shell scripts is very well possible, too, as I wrote above. Having spent a day or more debugging iSCSI booting via dracut, I actually started to hate people for using shell scripts for complex init stuff. Things get very messy there. Python would have been nice, but I guess that's to large for an initramfs. So there are not many options

As for being forced to use Linux when using systemd... Well, I'm wondering if some core stuff specific to linux could have been implemented using some plugin mechanism. But I perfectly understand the Lennards position of keeping the code clean and free from #ifdef hacks.

Then, nobody stops people from porting systemd to other kernels, keeping the APIs intact.

With systemd, there is one kernel and one way to start the system - and multiple distros.

Wouldn't it be nice if one could reuse the knowledge about init when switching distributions?*There is still package management and default desktops distributions can argue about