From fdabad92cedb32f1251e6c5a350d3bac0182899e Mon Sep 17 00:00:00 2001 From: Ivan Malison Date: Sun, 12 Jul 2026 13:20:17 -0700 Subject: [PATCH] docs: record K3s imagefs DiskPressure diagnosis --- dotfiles/agents/skills/disk-space-cleanup/SKILL.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/dotfiles/agents/skills/disk-space-cleanup/SKILL.md b/dotfiles/agents/skills/disk-space-cleanup/SKILL.md index f588c2fc..30abeb4a 100644 --- a/dotfiles/agents/skills/disk-space-cleanup/SKILL.md +++ b/dotfiles/agents/skills/disk-space-cleanup/SKILL.md @@ -125,6 +125,8 @@ Read `references/cleanup-patterns.md` before touching private service state, run Prefer service-aware sequences: stop the service, clean only validated cache/work paths, recreate required directories with correct ownership, restart, and verify health. +- 2026-07-10 `railbird-sf` incident: K3s reported `DiskPressure` even with tens of GiB free because its container `imagefs` shares `/`, kubelet image GC used the default 85% high-water mark, and the K3s config overrode only `nodefs` eviction thresholds. `crictl imagefsinfo` showed only ~677M of images, so image GC could not reclaim its requested ~149G and repeatedly evicted application pods. Set matching `imagefs.available` values alongside `nodefs.available` in `eviction-hard`, `eviction-soft`, and `eviction-soft-grace-period`; verify via K3s eviction-manager logs rather than trusting `df` alone. + ## Safety Rules - Do not delete user documents, source code, credentials, volumes, or unknown data without explicit approval.