[Nix-dev] NixOS 15.09 released

Alexander Zubkov green at msu.ru
Sat Oct 3 21:31:59 CEST 2015


Several days ago I found that PID 1 on my system permanently eats 30% of 
CPU by actively polling something. This is also somehow lead to some 
error when I have tried to do nixos-rebuild switch. I typed reboot and 
lost my machine. I found it powered off instead of rebooted when I 
reached it physically.

On 03.10.2015 22:02, Michael Raskin wrote:
>> Why don't you like systemd, guys?
>
> (I guess I have said everything below some time ago, when answering the
> same question)
>
> I would benefit from extermination of all software enforcing
> impossibility of anything similar to startx. Systemd is written so that
> it makes impossible working startx as a side effect. I want to write
> a small setuid binary starting a new X session (a second or a third
> one).
>
> I have an idea what I want my system to do and what I want my system not
> to do. Systemd spreads like cancer swiping more and more areas of the
> system into a single monolithic (even if multi-executable; it has no
> published stable interfaces) blob; each time it starts to do something,
> in a new area it takes additional effort to make it stop doing that, it
> is quite often very active by default. In some cases, like with session
> management, it is impossible to get back the simple behavior I always
> had and liked.
>
> I am still not sure if journald has reached the performance level of
> plain text logs. Also, apparently logs do get corrupted from time to
> time; with text logs I generally lose access to less data than with
> journald.
>
>>>> Oh, yes, another thing: independence of systemd :D
>>>
>>> Oh? I currently have some weird NixPkgs-based init-less system, how do
>>> I get NixOS services without systemd (or full NixOS without systemd as
>>> PID 1, I am OK with systemd running as a daemon if it doesn't maange
>>> logs and console sessions).
>
>
>
> _______________________________________________
> nix-dev mailing list
> nix-dev at lists.science.uu.nl
> http://lists.science.uu.nl/mailman/listinfo/nix-dev
>



More information about the nix-dev mailing list