Tumbleweed serves as my desktop, but for other purposes, I prefer a Leap-like system. I looked into other options and think I’ll go with Oracle 8.X because it has a longer lifecycle.
Considering one of the basic design goals for ALP is basically that, I don’t understand what all the FUD is about. It’s a different way of thinking about it and doing it, but nobody is taking away anybodies ability to run whatever the heck they might want on a system built with ALP principles.
1 Like
I installed a few of programs on MicroOS and rebooted numerous times, however the programs do not appear for normal users. Something is wrong if I have to use transaction-update shell to access a program that a normal user should have access to. The ALP concept is useful for anyone who frequently utilizes GUI software or maintains the system through a web-based interface. Aside from the DE, I create my own packages and am unable to use ALP.
Clear, you did give MicroOS a try and it was not a smooth experience.
AFAIK MicroOS will be the targeted to corporate users and Kalpa/Aeon the desktop variants of it.
Furthermore I did read above that Leap will be at least supported till end of the year 2025, 1.5 years (or more) from now.
So my conclusion is your tests were on pretty alpha software that will very likely be very different from what will become available.
@coldboot that’s what distrobox as your user is for, compile away…
I’ve had no issues with my own rpms, or installing other binary only packages for use by an end user…
A minor release of Leap (15.6) has been announced, which is beneficial. Let us see if the community decides to continue Leap on its own after 18 months.
I use distrobox to produce Debian packages from rpm spec files for clients, but it isn’t useful for anything else. For regular work, I require pfring, eBPF, DPDK, netflow recorder, and other tools that do not work well on the distrobox. My concern is that the ALP system will eventually cause harm to those who work in networks or security.
P.S.: Only time will tell if I am absolutely mistaken.
No, it won’t.
There are some things that need to be run in a “rootfull” distrobox distrobox create --root in order to have access to certain resources.
Is the documentation and/or workflow still a little up in the air on some of this? Certainly. But these are issues to be learned about and solved/documented.
Not issues that negate the entire concept.
Adding a broker and running a service over it increases latency and adds significant overhead.
Just finished installing openSUSE 15.5 today.
Took longer than expected, my older wireless app kept connecting then disconnecting, so no web checking or updates…
As sunset approached did move machine & monitor to where had a working cable then my connection worked fine, so ran sudo zypper refresh and sudo zypper up .
At least for today feel better…
Every dawn is a brand new year 
.
Didn’t they say something about a non-immutable openSUSE? They say they need voluenteers.
They want to build a Leap 16 but they need contributors.
Hmm, probably not my cup of tea. I read “a new approach to enterprise Linux for evolving use cases in a cloud-native world” and instantly feel I am not part of the target group.
I installed MicroOS in a VM, noticed that the root partition is read-only and dropped it. Then spent some hours to minimize the number of running services in Leap and am now happy with the result. Basically, Leap is the egg-laying wooly pig (“eierlegende Wollmilchsau”, as we say in German). I can use it for everything, from 1 GB 2 core mini VM to full blown desktop machine, private NAS and enterprise server. IMO, it is a huge competitive advantage to have the same code base everywhere.
Once I get the impression I am riding a dead horse, I’ll look elsewhere. Probably Arch Linux. Now that I learned that I rarely ever need to use Yast and can do everything from the command line, I’d feel comfortable going that route.
ALP is not the first approach to hide the underlying basics from the user. There have been numerous initiatives like that in the last decades and they all have one thing in common: they failed. Some people just don’t want to give away control and stay in control. Because they learned that automation only covers 80% of their needs and for the remaining 20% you have to dig deeper and are on your own. You. Need. Knowledge. Which comes from experience, doing. Mission statements like “Users should spend their time managing the application instead of managing the underlying Host.” may attract the technical buyers but rarely convince the technical staff and decision makers. My 2 cents.
1 Like
I just viewed this presentation:
And it looks to me the architecture of a Leap successor is far from being clear yet, “NewLeap”
1 Like
There is no roadmap or initiative that has been made public.
I tried MicroOS and after a few days grew to like it. I’m curious to see what happens when I compile and install programs in the traditional way.
You made some valid points. I’d rather have user control than an immutable system any day.
As Sfalken already pointed out several times: an immutable system does not mean to loose any user control. There is a missunderstanding what the concept and behaviour of an immutable system means…
2 Likes
As Sfalken already pointed out several times: an immutable system does not mean to loose any user control. There is a missunderstanding what the concept and behaviour of an immutable system means…
While an immutable system has some security advantages, it also means that users have less control over the system. Users may not be able to install new applications or configure the system as much as they would with a typical Linux distribution, for example.
I’ve gotten it to work to some extent, but I wish there were some options for overriding the immutable-bits and allowing the user to make appropriate adjustments.
Can you please explain what the meaning of “user(s)” is in this topic?
Is this meaning the person that “owns” the system, etc. and thus most probably is managing (installing and maintaining) the system and maybe also has a end-user UID for hios daily scores (banking, surfing the web, gaming, etc., etc.)>
Or is this the end-user (as in, not being the user root)?
1 Like
Henk pointed it out really well. “Normal end users” are not allowed to change the core of the system unless they use “su -” or “sudo” but then they are no longer “end user” but “administrator”.
“Normal end users” cannot change the base of the system.
“Admins” are able to change everything.
This applies for an immutable but also for an old fashioned distribution. So there is a basic missunderstanding of immutable systems…
Immutable systems add an additional security layer for the administrator that he doesn’t screw his system easily. He has a main running system and a copy where he aplies changes. If the changed copy boots well (additional logic checks are implemented) the copy gets the main system. If the copy wont work his system stays on the main. Thats a ruff and incomplete description of how it works…
1 Like