F5 patches five critical NGINX server security flaws

F5 has issued emergency security updates for five critical vulnerabilities in NGINX, warning that unauthenticated attackers could execute malicious code on affected web servers without possessing login credentials.

The flaws, disclosed on 18 November, affect NGINX Open Source and its commercial counterpart, NGINX Plus. Each carries a critical severity score of 9.2 under the Common Vulnerability Scoring System’s fourth edition, although successful exploitation depends on the server configuration and other technical conditions.

The vulnerabilities expose NGINX worker processes to memory corruption, potentially allowing attackers to disrupt websites or run unauthorised instructions with the privileges assigned to those processes. Full administrative control is not an automatic consequence of exploiting the bugs, security researchers cautioned.

F5 released fixes through an unscheduled security update, alongside patches for five additional vulnerabilities. The company has provided corrected NGINX Open Source versions 1.30.7 and 1.31.9, while affected NGINX Plus customers have been directed towards updated commercial releases.

The most exposed weakness, CVE-2026-123007, concerns the HTTP/2 protocol, used to handle communications between websites and browsers. It involves an integer overflow in flow control that can corrupt server memory when specially constructed requests are processed. A vulnerable server needs only to have HTTP/2 enabled for the flaw to become reachable.

Researchers at cloud security company Wiz demonstrated that the vulnerability could be exploited to execute code against a server delivering static web content. Their findings showed that the attack did not depend on a vulnerable application running behind NGINX or unusual third-party extensions.

Another HTTP/2 vulnerability, CVE-2026-123633, arises when a connection is closed while NGINX is communicating with an upstream server. The resulting use-after-free condition can cause software to access memory that has already been released, creating a further possibility of code execution or a worker crash.

The remaining vulnerabilities affect different features. CVE-2026-121810 involves HTTP/3 connections, while CVE-2026-123634 concerns the internal domain-name resolver in certain proxy configurations. CVE-2026-113425 affects installations using NGINX as a mail proxy with HTTP-based authentication.

All five flaws received scores of 8.1 under the earlier CVSS 3.1 framework. Their designation as critical reflects the newer scoring standard rather than evidence that every deployment can be compromised through the same technique.

The European Union’s computer emergency response team, CERT-EU, urged organisations to update affected installations promptly, giving priority to internet-facing systems. Its advisory stated that no exploitation of the five vulnerabilities in real-world attacks had been reported at the time of publication.

However, publicly available exploit code increases the urgency of patching. Security company Rapid7 demonstrated an attack against CVE-2026-123007 and released proof-of-concept material, while researchers described constraints involving memory protections and substantial volumes of network traffic.

Researchers at Xint grouped the vulnerabilities under the name NGINX2Root. They also described a separate vulnerability in the Linux kernel that could, under particular circumstances, be combined with NGINX code execution to obtain higher privileges. That kernel issue was not fixed by the NGINX updates and should not be confused with the five server vulnerabilities.

NGINX is deployed extensively as a web server, reverse proxy and traffic-handling component for online services. The software also appears within containerised applications and products maintained by other vendors, making it important for administrators to identify affected embedded installations rather than checking standalone servers alone.

Security company Sophos advised administrators to restart NGINX after installing updated packages so that the corrected executable replaces running code. It also noted that Linux distributors may incorporate security fixes into their own packages without adopting the same upstream version numbers.

Temporarily disabling HTTP/2 can mitigate the two vulnerabilities affecting that protocol, while disabling HTTP/3 can reduce exposure to the separate flaw involving that feature. Neither measure addresses the resolver or mail proxy weaknesses, and administrators still need the relevant security updates.

The affected versions span different periods because individual defects entered the software at different stages. The HTTP/2 flow-control weakness dates to NGINX version 1.9.5, released in 2015, whereas the HTTP/3 issue affects newer implementations. Operators therefore need to examine both the installed version and enabled modules before determining their exposure. A disabled or absent feature can prevent the corresponding attack route.

The AlmaLinux project said patched NGINX packages had been released directly to production repositories for supported software streams. Its maintainers identified affected packages across AlmaLinux 8, 9 and 10, while explaining that some retired software streams would not receive corresponding updates.



Notice an issue?

Arabian Post strives to deliver the most accurate and reliable information to its readers. If you believe you have identified an error or inconsistency in this article, please don't hesitate to contact our editorial team at editor[at]thearabianpost[dot]com. We are committed to promptly addressing any concerns and ensuring the highest level of journalistic integrity.


Loading next story…