Developers of the Clear Linux distribution have changed the project's development strategy. The primary focus is now on server and cloud systems, which will receive the main attention. Components for workstation editions will be maintained only as a secondary priority.
The delivery of desktop packages will continue, but these packages the original options for user environments, without Clear Linux-specific overlays or modifications. Packages with GNOME will continue to be formed, but the composition and settings of the desktop will match the reference version provided by the GNOME project by default.
The previously offered custom , separate , pre-installed third-party extensions for GNOME Shell (, , , ) and modified GNOME settings will initially be in a separate package "". Next week, packages with the desktop will be updated to GNOME 3.36, which will correspond to the reference GNOME environment, after which the package "desktop-assets-extras" will be marked as obsolete.
It is worth noting that the Clear Linux distribution is developed by Intel and ensures strict application isolation using containers divided through full virtualization. The base part of the distribution contains only the minimal set of tools for running containers and is updated atomically. All applications are packaged as Flatpak or sets (Bundles), run in separate containers. In addition to customized desktops, the developer edition is notable for extended hardware support, integration of a debugging system based on FUSE, the addition of a new installer, and the presence of , which offers sets for deploying development environments using various programming languages and technologies.
Features of Clear Linux:
- The binary delivery model of the distribution. System updates can be performed in two modes: applying patches to the running system and a complete system update through the installation of a new image in a separate Btrfs snapshot and replacing the active snapshot with the new one;
- Aggregation of packages into bundles (), forming ready functionality, regardless of how many software components they consist of. The bundles and system environment images are created based on an RPM package repository, but are delivered without package separation. Inside the containers, a specially optimized copy of Clear Linux is executed, containing the required sets (bundles) to launch the target application;
- An efficient update installation system, embedded into the core of the distribution, ensuring accelerated delivery of updates that fix critical issues and vulnerabilities. An update in Clear Linux includes only the data that has changed, so typical patches for vulnerabilities and errors occupy only a few kilobytes and are installed nearly instantly;
- A unified versioning system — the version of the distribution represents the state and versions of all its components, which is convenient for creating reproducible configurations and tracking changes in the distribution's components at the file level. Changes/updates to any part of the system always result in the overall version of the distribution being changed (while in conventional distributions only the version number of a specific package increases, in Clear Linux the version of the entire distribution increases);
- A stateless approach to configuration definition, suggesting that different classes of settings are separated (OS, user, and system settings are stored separately), the system does not maintain its state (stateless) and does not contain any settings in the /etc directory after installation, generating settings on the fly based on templates specified at startup. To reset system settings to default values, one can simply delete /etc and /var;
- full virtualization (KVM) for running containers, which ensures a high level of security. The startup time for containers is slightly behind traditional container isolation systems (namespaces, cgroups) and is acceptable for demand-based application container launches (the virtual environment startup time is about 200ms, with an additional memory consumption of 18-20 MB per container). To reduce memory consumption, a mechanism is implemented (direct access to the file system bypassing page cache without using block device level), while the technology is used for deduplication of identical memory areas (Kernel Shared Memory), which allows for the organization of resource sharing of the host system and connects the same system environment template to different guest systems.
Source: opennet.ru
