Systemd Rolls Out Its Own Mount Tool

[anda_skoa]

Linux desktop users generally need systemd less than hog needs tits, some don't care, some are against it, some are supporting.

Desktops users don't care or think they don't need it because they are at least one more level of tooling away.

I.e. systemd is relevant for the developers creating the software these desktop users use.

They would care if they knew that features they have have only been possible because of systemd or have made it easier for the desktop developers and freed up resource for other features, etc.

As a quite similar situation take file systems.
Sysadmins will care directly, because they need to deal with them directly.
Desktop users don't, but the developers of desktop software might not be able to provide certain features on some file systems or not provide them as easily or consistently, etc.

And it will also not be the same for all desktop software developers, since different types of desktop software have different requirements.
E.g. a notes taking application might not care whether the file system is ext4 or vfat, a video editor might not like 4GB limits, a backup tool might require hardlink/symlink capability.

Similar for systemd, developers dealing with management of background processes, e.g. desktop shell developers, might care a lot, developers of system management tools as well, developers of games much less so.

In either case the users are too far away from these details

Cheers,
_

[sdack]

As per usual, the pro-systemd crowd have absolutely no idea what they are doing. Makes for sad threads though!

Fixed it.

[duby229]

Desktops users don't care or think they don't need it because they are at least one more level of tooling away.

I.e. systemd is relevant for the developers creating the software these desktop users use.

They would care if they knew that features they have have only been possible because of systemd or have made it easier for the desktop developers and freed up resource for other features, etc.

As a quite similar situation take file systems.
Sysadmins will care directly, because they need to deal with them directly.
Desktop users don't, but the developers of desktop software might not be able to provide certain features on some file systems or not provide them as easily or consistently, etc.

And it will also not be the same for all desktop software developers, since different types of desktop software have different requirements.
E.g. a notes taking application might not care whether the file system is ext4 or vfat, a video editor might not like 4GB limits, a backup tool might require hardlink/symlink capability.

Similar for systemd, developers dealing with management of background processes, e.g. desktop shell developers, might care a lot, developers of system management tools as well, developers of games much less so.

In either case the users are too far away from these details

Cheers,
_

I get it and to a very limited extent I do agree with you, but there is more imagination to your argument than reality. Really what your argument boils down to is the distributions target audience. If the distro targets non computer literate people then I'd say you are largely correct, but for every other target you are definitely not.

It really sucks how many distros adopted systemd, there used to be tens of desktop linux distributions that computer literate enthusiast class users could choose from, now there are a whole lot less. That's systemd's fault. If their goal was to kill linux as an enthusiast platform then they have been largely successful.

[anda_skoa]

I get it and to a very limited extent I do agree with you, but there is more imagination to your argument than reality. Really what your argument boils down to is the distributions target audience. If the distro targets non computer literate people then I'd say you are largely correct, but for every other target you are definitely not.

I can't agree on this one, there are just to many levels of computer literacy and even within different requirements for people within the same level.

E.g. while you might have the knowledge to do low-level sysadmin things you might not want to or not want on all your systems.
I've configured my own firewall devices in the late 90's and early 0's, even special hardware without out GUI and jsut serial console, yet I don't need that kind of personalized/controlled setup on my laptop.
However, at the same time I also know a couple of friends who do their laptop network setup with hand crafted scripts.

But in general I would say that it is fair to assume that deskop users don't really care if the network is setup by some script or NetworkManager or systemd, but that they would care if they could not log on to a hotel WLAN through their usual desktop applet.

E.g. I would care despite being able to handle my Debian system's network config, wpa_supplicant, etc.

It really sucks how many distros adopted systemd, there used to be tens of desktop linux distributions that computer literate enthusiast class users could choose from, now there are a whole lot less.

Can you give an example of effects caused by the change.

I am using Debian (unstable) and I didn't get the impression that I was in anyway limited. I could still build my own kernel (never actually needed that), can still do fine grained installation (always used net-inst based minimal installs and then add only what I need), can still select file systems, lvm layouts, crypto setups, as always.

For me the only change I am regularily aware of that I look at the system log by running "journalctl -f" instead of "tail -f /var/log/syslog"
(in either case part of my shell history obviously, so the long term change is jsut the necessary characters to type until completion succeeds)

Cheers,
_

[trek]

I like systemd in concept, but implemntation has...

Or take static IP, for example. How do you persuade system to change setting on the NIC without rebooting it ?

Or mounting. I wanted to have NFS4 mount on a map that would be hardlinked to map2, so that I could have alternate local content in case nfs4 mount fails.

I coulnd't persuade systemd to do mounts in desired order, so hardlink would always happen AFTER NFS4 mount.

You described a completely unmanageable system, congrats to systemd!

[trek]

It seems that lots of the systemd haters never really used it yet. Maybe you're clutching your old Debian version and hoping that EOL will never appear that you have switch to systemd.

I'm using the latest Debian and I have not systemd installed. How it is possible? Well, I just switched to the not-installed-by-default sysvinit-core package! No need to change my distro

[gilboa]

why develop systemd can't we just stole svchost from MS?

You have no idea what you're talking about.
systemd and svchost has *nothing* in common.

Repeat after me:
systemd is a set of Linux base services. Its main aim to is replace *all* the difference tools and utilities that make Linux base-system tick.
Its not (only) an init system (initrd).
Its not (only) a service wrapper ('svchost').

And I for one, as someone how dropped close to 5,000 LOC when I dropped initrd and moved to to systemd, welcome my new Linux based system overlord.

[Passso]

My opinion about systemd changed when I discovered that :

systemd-analyze blame

Now I like it I do know exactly what is going wrong in my system. Before I had to search for hours which service was slowing the system boot.

[sdack]

Desktops users don't care or think they don't need it because they are at least one more level of tooling away.

I.e. systemd is relevant for the developers creating the software these desktop users use.

They would care if they knew that features they have have only been possible because of systemd or have made it easier for the desktop developers and freed up resource for other features, etc.

As a quite similar situation take file systems.
Sysadmins will care directly, because they need to deal with them directly.
Desktop users don't, but the developers of desktop software might not be able to provide certain features on some file systems or not provide them as easily or consistently, etc.

And it will also not be the same for all desktop software developers, since different types of desktop software have different requirements.
E.g. a notes taking application might not care whether the file system is ext4 or vfat, a video editor might not like 4GB limits, a backup tool might require hardlink/symlink capability.

Similar for systemd, developers dealing with management of background processes, e.g. desktop shell developers, might care a lot, developers of system management tools as well, developers of games much less so.

In either case the users are too far away from these details

Cheers,
_

systemd addresses mostly symptoms. It doesn't solve the causes. init was simple, but it wasn't bad. Only did it not fix the symptoms, nor was it ever designed to do this, because the ideology was to create a coherent system with every developer only adding more coherency. Now people are trying to sell systemd as good and want to label init as bad, because they don't know how to solve their problems otherwise.

We ended up in this situation, because distributions have raced each other in the amount of supported packages and given little care to the consequences. It has turned from a once coherent operating system into an incoherent conglomerate. We are less often fixing the causes and more often mending the symptons. systemd isn't bad in itself, nor was the init system, but it is the approach the maintainers have taken, which is bad. Bad, because now we have a tool for mending a ton symptoms without a need to solve the causes. Today it is the systemd-mount, tomorrow it might be a systemd-su, then a systemd-chown and who knows how far it will go. God knows if we might end up with a systemd-registry...

If you only ever want to address symptoms then systemd is the future. What people hate is the loss of an ideology (that of having coherency across all software). They hate the rise of this "Master Control Program" and they hate the increased complexity of a system, which they got to know as one to be be simple, coherent, transparent and easy to understand. And they hate the loss in quality, because systemd serves the quantity of packages, but not necessarily the quality of these.

Again, I am not complaining (more like trying to create a bridge). I have seen it it before, but I am not going to support the idea of this being a positive development. It is more like accepting an evil - like accepting Windows.

[stefansaraev]

It really sucks how many distros adopted systemd, there used to be tens of desktop linux distributions that computer literate enthusiast class users could choose from, now there are a whole lot less. That's systemd's fault. If their goal was to kill linux as an enthusiast platform then they have been largely successful.

shocking. how can that be systemd fault?

oh right, it's always everyone else's fault. never yours.

and btw, you still have dozens of systemd-free distributions.

for you, linux may be an "enthusiast" platform, for many of us, it was never like that.