Repository navigation
kindnet: bump kindnetd to v20260820-69b56db7 - #239
Conversation
The pinned kindnetd digest dates from 2025-05-12 and is now 15 months old. generate_manifests.sh always resolves the latest tag from Docker Hub, so this is simply the result of re-running it; the assets had not been regenerated since the original pin. This clears two CRITICAL findings in the kindnetd binary itself: CVE-2026-33186 google.golang.org/grpc v1.72.0 -> fixed (authz bypass, 9.1) CVE-2025-68121 Go stdlib 1.24.2 -> built with Go 1.26.7 Two CRITICALs remain in the new image, both outside bin/kindnetd and not fixable here: CVE-2025-68121 in the go-runner helper (Go 1.25.6) and CVE-2026-31789 in the Debian base (libssl3 3.0.18, fix lands in 3.0.19). The DaemonSet passes no args or command and configures kindnetd purely through HOST_IP, POD_IP and POD_SUBNET. Both binaries were extracted and compared: entrypoint (/bin/kindnetd), the CLI surface (logging flags only) and all three environment variables are unchanged, so the bump is drop-in. POD_SUBNET in particular is still honoured, which matters because getting it wrong is what breaks source-IP preservation (see microshift-io#236). Only the kindnet assets are regenerated here. Re-running the script also bumps kube-proxy to v1.37.0, which is left out on purpose: that image has to track the cluster's Kubernetes version rather than the newest published tag.
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Team Run ID: 📒 Files selected for processing (4)
Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review. 📝 WalkthroughWalkthroughThe pull request updates pinned kindnet container image digests in architecture-specific kustomization and release asset files for aarch64 and x86_64. ChangesKindnet image digest updates
Estimated code review effort: 1 (Trivial) | ~2 minutes Merge Risk: ⚪ Minimal · up to This updates the pinned kindnet image digests consistently across release assets for both supported architectures. No current merge-blocking risk remains. Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Full details: Docstring CoverageExplanation No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0 files. (4 skipped: 4 unsupported.) ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
The pinned kindnetd digest dates from 2025-05-12 and is now 15 months old.
generate_manifests.shalways resolves the latest tag from Docker Hub, so thisis simply the result of re-running it — the assets had not been regenerated
since the original pin.
What this fixes
Two CRITICAL findings in the kindnetd binary itself:
Two CRITICALs remain in the new image, both outside
bin/kindnetdand notfixable here: CVE-2025-68121 in the
go-runnerhelper (Go 1.25.6) andCVE-2026-31789 in the Debian base (libssl3 3.0.18, fix lands in 3.0.19).
Why it is drop-in
The DaemonSet passes no
argsand nocommand; kindnetd is configured purelythrough
HOST_IP,POD_IPandPOD_SUBNET. I extracted both binaries andcompared them: same entrypoint (
/bin/kindnetd), same CLI surface (loggingflags only), and all three environment variables present in both.
POD_SUBNETmatters in particular, because getting it wrong is what breakssource-IP preservation behind the router (see #236).
Verified on a running cluster
Installed on MicroShift 4.22 (Kubernetes v1.35.6), single node, kindnet as CNI:
POD_SUBNETstays10.42.0.0/16on both sidesRunning, 0 restarts; 53/53 pods healthy after the restartKIND-MASQ-AGENTstill carries theRETURNfor the pod CIDR, and realexternal client IPs continue to arrive in
X-Forwarded-For— so source-IPpreservation is intact
Scope
Only the kindnet assets are regenerated. Re-running the script also bumps
kube-proxy to v1.37.0, which is deliberately left out: that image has to track
the cluster's Kubernetes version rather than the newest published tag, and on
a 1.35 cluster it would be two minor versions ahead.
Summary by CodeRabbit