Hi guys. I have conky running and it shows swap usage. I deactivated swap with "swapoff -a", after which conky started to show "no swap" message. Then I loaded a lot of YouTube videos which took more that a half of my physical RAM and the message in conky disappeared. I ran "swapoff -a" again and it appeared again. Did my system activate swap without my command?

On systems with sysvinit this shouldn't be the case, swap is only mounted once automatically, at boot time. Since Arch uses systemd there may be some automatism that automounts swap when some event is triggered or just in a specific interval. I would start to check the cron tables for suspicious actions of that kind.

Guess I have nothing in my crontab. Must be systemd behaves the Windows way.

1 members found this post helpful.

Thanks, syg00. "deleting the systemd-gpt-auto-generator, and adding its path to the NoExtract rule for pacman" looks like the best way to get rid of this behaviour. But there's no such package. Systemd is very complicated for me, can you please give me a hint on how to cut it out of the system? Just looking at editing units gives me a headache.

I haven't looked at Arch since they went to systemd, but the wiki is always excellent. I migh be inclined to use "systemctl | grep -i swap", then just use systemctl to stop the relevant unit.

If indeed systemd is activating swap space that explicitly was manually disabled by root then I would consider that to be a bug and you should file a report upstream.

The whole systemd is a bug and should be reported. But devs like bugs these days.

If you don't like systemd it may be time to think about switching distros.

Maybe I'm just slow on the uptake, but if you have a swap partition, why disable it? And if you don't want swapping, why have a swap partition?

Are there any that won't ever switch to systemd?

I have swap because several days ago I had just 2 GB of RAM. Now 4, and I was just playing with swap command. And noticed this behaviour during experimenting. I know I can remove swap partition but that's another question. I saw how my Linux box did something out of line, which I didn't expect.

No one can say, maybe systemd is at some point inevitable, maybe (hopefully) it will fail before that before that point is reached. Don't get me wrong, I don't wish systemd to fail, if it would be just another init system I couldn't care less, but I think its IMHO cancerous nature will at one point bring it down. We can only hope that it doesn't do too much damage to the Linux ecosystem.

However, there are definitely distros that will avoid it as long as possible (like Slackware or CRUX) or that have it in their repositories, but offer alternatives (like Gentoo).

The problem with systemd spreading very quickly is that more and more devs begin to write their software in a way when it interacts with systemd. Gentoo maintainers try to avoid this evil and they already have to fork some programs (what was it, "udev"?). If more and more programs will be written to work with systemd, at some point there will be no possible way for distros' maintainers to keep up with forks. If I think about it, I don't believe Gentoo won't give up and just go with the flow and switch to systemd in next 5 years.

Most programs don't have anything to do with systemd, luckily, it is mostly DEs/WMs that have to deal with it. The average file-manager, image manipulation software, text-editor, browser and other common software do not have to rely on systemd and it is likely that especially the larger programs for the sake of being platform independent will refuse to rely on it. The part where it might become gard to avoid systemd is that what is happening under the hood (booting, managing services, power-management, detecting and configuring hardware). We will see what happens.