Recently I read an article in How-to-Geek about changing from a resource-hungry distro (Ubuntu) to a resource-thrifty one (Manjaro) in order to speed up processing and require less support from a resource-challenged computer. It included this bash command to prove its point:

systemctl list-unit-files --state=enabled --no-pager

The article pointed out that the heavyweight distro had 90 running daemons when idle, while Manjaro had only 24. I've used it as part of my evaluation of various distros I might adopt for everyday use, including on a small laptop that I travel with.

Someone should tell the author about Solus. I just ran that command on a Solus Budgie VM that I'm currently running, and the result was 12! That suggests that the team has done a wonderful job of cutting out the unnecessary waste.

For anyone who's interested, here's the article: https://www.howtogeek.com/430556/why-i-switched-from-ubuntu-to-manjaro-linux/

I got 14 on mine, still prety good πŸ˜‰

I found this to be a very interesting article. The author also praised manjaro's big 3rd Pty repo and it's AUR. I would think heavy employment of these things (neither which I'm interested in) would skew results but maybe not dramatically. I could be wrong there...Nonetheless, I think being a resource hog is a great criteria on judging an operating system and service daemons will do that.
I had 20. 8 are snap-related which is baffling considering I never use it...still good though.

    Today I'm using Solus 4.0 Budgie on a real computer, not a VM, and the same command reported 13 daemons. It's not just a fresh install, like those in the article - it's fully functional. Tomorrow morning I'm going to replace Manjaro on our multimedia computer with Solus. I can use all the speed I can get on that machine - it runs Netflix and plays Blu-ray videos from my NAS server.

    brent Installing tons of applications don't impact the performances as long as they are not running... It just consume more disk space. The problem with AUR is that you can find anything there: good stuffs but also unmaintained things, badly packaged applications, packages abandoned by their maintainer or not updated at the same rhythm as the core stack resulting in breakages, it already happened that some package contained malwares or crypto miners, it's not always clear for the end user if the package is built from source or if it's a binary that is packaged, experimental/unstable things are also available there, some package are illegal, etc. So it might be handy but it might be awful as well. People who use AUR should check the packages.

    Regarding the number of services, well having useless services running isn't a good thing, but just focusing on the number of enabled units is just stupid !
    A very obvious example of less doesn't mean better : stop the networkmanager service and you won't have network/internet connectivity anymore !
    Also some units are targets : their are used for grouping units, sockets are just listening on IP/ports but aren't daemons as such and timers are the equivalent of cron/scheduled tasks -> if you schedule a task to run every 30 days, it will use resources only when it's running not all the time and it could be for very useful things like performing logs cleaning, trimming a SSD drive, backups, etc...

    So yeah Ubuntu enable lot of things, it's a choice they make to have a maximum amount of things working out of the box and thus offering a great user experience. This of course bloat the system and distos like Solus don't want that and try to offer a good user experience without enabling tons of stuffs "in case somebody would need it".
    -> My criticism is that 90 vs 24 is an indicator but it's completely meaningless since absolutely no analysis of what is loaded and why.

      kyrios Not only do I appreciate that you give me a great education when I misinform myself, but you know I want to be corrected when I am misinformed. I never want to stop learning about linux/solus.
      That said if the numbers of processes mean nothing without analysis, then the numbers (90 v 24) still would be an indicator of resources used, though?

        brent Not really. It makes much more difference if the processes take up CPU or RAM than how many there are. How many doesn't mean much, they may be mostly sleeping/idle/swapped out.

          dbarron That makes sense, thank you.
          I got lulled into thinking there is a quantifiable correlation by the article author's own words:

          "So, what can explain the speed benefits? As an example, let’s look at the number of services and daemons that are running by default. They each consume system resources, such as a little memory and some kernel time. You can check on the enabled services and daemons by typing the following command in a terminal window...These are two fresh installs. As you can see, Manjaro has 24 enabled daemons, and Ubuntu has 90. That kind of overhead cannot fail to have an impact."

          Let me re-phrase that. I got lulled into thinking this was so in my own interpretation of his words---no disparagement intended at all to the credentialed article writer.

          kyrios

          that kinda reminds of those windows xp "tuning" days, lots of people made their "lite" versions of windows xp where almost all background services were disabled. Then they didn't understand why simple things didn't work anymore after that. I had to deal with this stupid stuff a lot back in the days, but was kinda funny too.

          Anyway, I've 12 unit files listed on my Solus. lol.

            Here's more than I ever expected to know about units and unit files:

            https://www.digitalocean.com/community/tutorials/understanding-systemd-units-and-unit-files

            I think the original author intended the --state=enabled option to separate the active ones from all the rest. If you run the systemctl list-unit-files with the --all option, the list includes hundreds of units, but very few of them have their state set to enabled. Most are disabled or static.

            It would be interesting to compare the list of 90 from Ubuntu with the list of 12 from my Solus VM in order to see just what all the bloat represents. Based on the way Solus performs for me, I'm pretty sure nothing important is missing. πŸ˜„

            21 unit files listed.

            A good chunk of them are snap-related too.

            nodq That was me 2000/XP/W7....used a disable guide for years from a guy called Black Viper online. I didn't do it for speed, I did it for security. Never broke nothing though, ever.

              Just tried this on my Debian install which has 60 running compared to 12 on my Solus machine. AppArmor is running on Debian but not on Solus and it is installed.I don't know much about how AppArmor works but are profiles not loaded on Solus the way it is loaded on Debian?

                onthebeach I've never used AppArmor, but the rest of your message makes sense to me. Ubuntu is based on Debian, if I remember right, so it makes sense that Debian has 60 and Ubuntu has 90. Of course, it might depend partly on the desktop environment, too.

                  brent yeh then you were one of those few people that knew what you did there. But too many didn't know what they do, they just followed some stuff they found in magazines or online. Then complained how simple things did not work anymore. It was like 3 times a week I could check those windows xp pcs of people doing this shit lol.

                    WetGeek I should have mentionted that both were not a fresh install, so the list for Debian likely less than 60. I'm counting at least 7 (maybe 11 if the other 4 are for virtualization) that were directly from what I installed. I'm using Gnome on both systems and use Flatpaks instead of Snap. I'm also using the 'unstable' repository on Debian.

                    I've 12 unit files listed on my Solus, and 69 unit files on my LinuxMint