Linux Kernel Virtual Memory & Cgroups v2
Understanding how the Linux kernel allocates physical pages via the Buddy Allocator and handles memory pressure with the Out-Of-Memory (OOM) killer is crucial for Kubernetes node stability.
Cgroups v1 vs v2 Unified Hierarchy
In cgroups v2, process management uses a unified tree structure located at /sys/fs/cgroup. Unlike v1, controller limits are configured via high, min, and low memory thresholds:
# Inspect unified cgroup memory metrics
$ cat /sys/fs/cgroup/memory.current
$ cat /sys/fs/cgroup/memory.high
$ cat /sys/fs/cgroup/memory.max