Fedora CoreOS To Enable systemd-oomd & zRAM Swap By Default

Fedora CoreOS is adopting two memory management changes that reflect broader trends in container and cloud-native environments. systemd-oomd moves OOM decisions from the kernel to user space, allowing for more granular control over which processes get killed under memory pressure. This is particularly relevant for containerized workloads, where the kernel's traditional OOM killer may not account for cgroup hierarchies or application priorities. The shift means engineers will need to understand systemd-oomd's policies and may need to adjust them for workloads that previously relied on kernel behavior.

Enabling zRAM swap by default addresses a common pain point in memory-constrained environments: slow disk-based swap. zRAM compresses memory pages in RAM, effectively increasing available memory without adding physical RAM or incurring disk I/O latency. For Fedora CoreOS users, this means better performance under memory pressure, but it also introduces a CPU overhead for compression/decompression. The change is most beneficial for systems with limited RAM, but engineers should monitor CPU usage if their workloads are already CPU-bound or if they're running on low-power hardware.

The combination of systemd-oomd and zRAM swap represents a significant shift in how Fedora CoreOS handles memory management out of the box. While these features have been available in other distributions, enabling them by default signals a move toward more proactive memory management in container-optimized OSes. Engineers deploying Fedora CoreOS will need to validate their workloads under these new defaults, as the changes could alter application behavior under memory pressure. The lack of explicit configuration means these features will be active even in existing deployments, requiring teams to understand their implications rather than opt into them.