Some readers have noticed that I seldom write in this blog – at least compared to certain periods before 2023. One reason for this has been and is: Linux – in my case Opensuse Leap 15.x – has just worked well for me up to this very day. Small glitches, but on average Leap 15 and its subversions have always fulfilled my needs and requirements. (Another reason is that after retirement, I got more interested in math, physics, ML and politics.)

However, I also have to say that my private needs have become very limited. Significantly fewer server components due to a removal of LAMP/Postgre-SQL/JS/JQuery and Kolab stacks and keeping only a minimum of LDAP services. Only some virtualization for server-based mail services. Actually, I have become a desktop user, sometimes using a Python/CUDA stack for local Machine Learning experiments. Aside of mail services, my own servers have become simple data containers. And the few requirements some remaining customers of my wife have with respect to their web-installations do not require full-fledged development stacks.

Now, as a pure desktop user, I am approaching a point (EOL of Leap 15.6) at which I seriously have to decide whether I should turn to Opensuse Leap 16 or move to a completely different distribution as Debian.

The reason is that Leap 16 comes with a lot of fundamental changes and a significant reduction of supported application packages. A new installer (Agama), no Yast any longer, the container-based ALP-system mixed with RPMs, confusion about immutability and BTRFS, Myrlyn as an interface to package management, SE-Linux instead of Apparmor, no support for older Nvidia cards, Agama installer (according to some users) not fully working with Luks and LVM, … Well, what I read smells like a bunch of potential problems.

Regarding ALP and containers, I have used some applications like Blender for a long while now with Snap and/or Flatpak – and was always shocked about the additional disk space requirements. So, I am very skeptical regarding this aspect. It may be suited for Cloud based services – but for local installations in an environment with a mixture of desktop systems and small servers? The question is whether and in which form ALP affects small file server installations.

Aside of ALP, what is my present strategy regarding my desktops and laptops?

I will give Leap 16 a chance. As a fist step, I will install it freshly on an external USB disks for my various systems and see what the installation leads to. I will start with a mutable version for desktop usage. One objective is to install Leap 16 with multiple partitions, all encrypted with LUKS and, of course, using LVM. For desktop systems with relatively new Nvidia cards I expect full support (including CUDA).

As for an old Optimus laptop which has an Nvidia card that requires the version 470 Nvidia drivers, I will probably have to accept that its card is no longer supported. So, I will have to see how far one gets on such a system with the Nouveau driver and a Noveau-adapted version of prime-select. I hope that Blender runs with reasonable performance even on these systems. If everything works as expected, I will start with an upgrade of the existing 15.6 installations on desktops and laptops.

If I, however, run into serious trouble, I will definitely switch to Debian. Debian is my preferred alternative, based on very good experiences with Debian on servers and Kali Linux on KVM virtualized desktops.

The decision to keep Leap or not may depend on the kind of system and their age.

I will report about my experiences during February. So, stay tuned …