DEV Community

Cover image for NGINX2Root: CVE-2026-123007, HTTP/2 RCE & Patching
Naveed Ahmed
Naveed Ahmed

Posted on Originally published at blog.naveedkumbhar.com

NGINX2Root: CVE-2026-123007, HTTP/2 RCE & Patching

Originally published on Naveed Ahmed's Engineering Blog.

NGINX is often the first application component that receives traffic from the public internet.

It's used for reverse proxies, API gateways, containerized web applications, and Kubernetes ingress.

That makes NGINX security updates especially important for infrastructure teams.

On November 18, 2026, several NGINX vulnerabilities were disclosed under the name NGINX2Root.

The principal issue is CVE-2026-123007, a memory-corruption vulnerability in HTTP/2 flow control.

Successful exploitation may allow unauthenticated code execution within the NGINX worker-process context.

Here's a practical defensive overview.

1. Understand the Five Vulnerabilities

CVE Component
CVE-2026-123007 HTTP/2 flow-control memory corruption
CVE-2026-123633 HTTP/2 connection-draining use-after-free
CVE-2026-123634 Dynamic resolver use-after-free
CVE-2026-121810 HTTP/3 use-after-free
CVE-2026-113425 Mail-proxy use-after-free

Each vulnerability has different feature prerequisites.

The full NGINX2Root exploit demonstrations involve more than just compromising an HTTP/2 worker; escalation to host root requires an additional vulnerability and suitable conditions.

2. Check Your NGINX Version

On a standalone host:

nginx -v
Enter fullscreen mode Exit fullscreen mode

Review compiled features:

nginx -V
Enter fullscreen mode Exit fullscreen mode

Inspect the active configuration:

sudo nginx -T
Enter fullscreen mode Exit fullscreen mode

The patched upstream Open Source releases are:

  • Stable: 1.30.7
  • Mainline: 1.31.9

For distribution packages, check vendor advisories rather than relying on the displayed upstream version alone. Maintainers may backport security fixes without changing the upstream version number.

3. Review HTTP/2 Exposure

Search the effective configuration for HTTP/2-enabled listeners.

For example:

sudo nginx -T 2>&1 | grep -E 'http2[[:space:]]+on|listen .*http2'
Enter fullscreen mode Exit fullscreen mode

The result needs manual interpretation because generated includes and deployment-specific configuration may change what is active.

For vulnerable HTTP/2 listeners, temporarily disabling HTTP/2 can mitigate CVE-2026-123007 and CVE-2026-123633.

For supported configurations, an example is:

server {
    listen 443 ssl;
    http2 off;

    server_name example.com;
}
Enter fullscreen mode Exit fullscreen mode

Validate and apply the configuration using your normal change-management process.

Disabling HTTP/2 is a temporary mitigation, not a substitute for patching all affected features.

4. Audit Kubernetes Ingress Controllers

First, identify which ingress implementations are actually installed.

kubectl get deployments,daemonsets -A
Enter fullscreen mode Exit fullscreen mode

For the community ingress-nginx controller, inspect its Pods and images where the standard labels are present:

kubectl get pods -A \
  -l app.kubernetes.io/name=ingress-nginx \
  -o wide
Enter fullscreen mode Exit fullscreen mode

Review the exact image reference and image digest.

Then determine which NGINX binary is bundled inside that image and check the controller vendor's security advisory.

Do not assume a specific ingress controller version is safe merely because its version looks recent.

Community ingress-nginx and F5 NGINX Ingress Controller are different products with separate release schedules.

5. Apply and Verify the Security Fix

On a Linux host, follow the distribution vendor's update process.

After updating, ensure affected running processes have been replaced as required by the vendor guidance.

A graceful configuration reload does not necessarily replace the executable image used by the existing master process.

A controlled restart may be needed:

sudo nginx -t
sudo systemctl restart nginx
sudo systemctl status nginx
Enter fullscreen mode Exit fullscreen mode

Restarting can affect active connections. Use a load-balanced rollout or an approved maintenance procedure where availability matters.

For Kubernetes, update the controller image to a vendor-confirmed fixed build and roll it out using an availability-aware deployment process.

Verify traffic and running images after deployment.

6. Review Compromise Indicators

Check for unexpected worker failures, suspicious child processes, unusual outbound connections, and security alerts.

A worker crash alone is not proof of successful code execution.

If compromise is suspected, preserve evidence and follow your organization's incident response procedure.


Read the Complete NGINX2Root Guide

My full article explores the HTTP/2 vulnerability, companion CVEs, patching considerations, and a Kubernetes ingress triage runbook.

Read NGINX2Root: HTTP/2 RCE & Resolver Vulnerabilities

Official references:


How does your team verify that all externally exposed NGINX deployments are patched?

Written by Naveed Ahmed — DevOps & Platform Engineering

Top comments (1)

Collapse
 
onelifevibe profile image
OneLifeVibe •

The distinction between community ingress-nginx and the F5 controller is the part that will save teams the most time here, since a version check alone can look reassuring while the bundled binary is still vulnerable. The reminder that a graceful reload does not replace the running executable is equally important, and it is easy to miss during a hurried patch window. For the HTTP/2 exposure check, did you find the nginx -T grep reliable in environments with generated includes, or did you need a second pass against the rendered config to catch listeners defined in included files?