2026-10-01 –, Galerie
Delivering a Linux OS via OCI images unifies infrastructure management but often introduces severe network and state-reconciliation bottlenecks. This talk dives into the mechanics of solving these challenges by applying OS updates directly from OCI images to btrfs snapshots. We will explore how to drastically optimize these atomic upgrades using zstd-chunked compression, enabling systems to execute file-level delta updates via HTTP range requests instead of pulling full layers. Finally, we will demonstrate how to leverage snapper for seamless 3-way merges, safely reconciling upstream OS changes with local user configurations.
The transition to image-based, immutable Linux systems fundamentally changes OS management, but it also introduces complex network and storage challenges that are often abstracted away. This session explores the architecture and technical workflows required to provision and upgrade an OS using OCI images, advanced compression, and copy-on-write filesystems.
The foundation of this approach relies on mapping OCI images directly to btrfs snapshots. OCI image layers are pulled, unpacked, and directly applied as immutable btrfs snapshots, forming the read-only root volume of the system while maintaining distinct writable areas for persistent state.
To eliminate the network bottleneck of pulling full OCI layers for minor patches, we will detail efficient delta updates with zstd-chunked OCI images. The OS agent identifies and downloads only the specific files that have changed between the current running image and the target image. We will explain the mechanics of using HTTP range requests to extract these payloads and write them directly into the target btrfs snapshot on the fly, drastically reducing update times and bandwidth consumption without requiring full local extraction.
Finally, an immutable OS still requires localized configuration (e.g., in /etc), creating a challenge when the base OS image is updated. We will explore how to solve this using a 3-way merge strategy. In particular we will see how snapper is utilized to navigate btrfs snapshot trees and isolate the precise differences between the original base image, the user's localized state, and the incoming update. Having a clear picture of the differences in multiple directions allows us to define an algorithmic approach to safely merging these three states.
Starting his career at a research foundation, David Cassany contributed to projects focused on multimedia streaming within open-source Linux ecosystems. For the past decade, he has brought that expertise to SUSE as an open-source software engineer, specializing in Linux appliance build tools, containerized systems, and Kubernetes architectures.
Dirk is a Distinguished Engineer, working on many container and operating system related topics at SUSE.