NGINX HTTP/2 flaw can lead to unauthenticated code execution: fixes ship in 1.31.9 and 1.30.7
On November 18, 2026, NGINX released versions 1.31.9 and 1.30.7 to fix CVE-2026-123007, a memory corruption flaw in HTTP/2 flow control. It has been in the code since 1.9.5, and F5 says an attacker can run code where ASLR is off or can be bypassed. The Hacker News reports no attacks so far.
NGINX shipped fixed versions for an HTTP/2 flaw
With versions 1.31.9 and 1.30.7, dated November 18, 2026, NGINX fixed CVE-2026-123007. The nginx.org security page lists it as memory corruption in HTTP/2 flow control and rates its severity as major. The affected range runs from 1.9.5 through 1.31.8, while 1.31.9 and later, and 1.30.7 and later, are not affected.
The release notes say the bug is an overflow of the HTTP/2 flow control window that can crash a worker process or have other impact, and they credit Xint.io. The Hacker News adds that the flaw has been nicknamed nginx2root and has been in the code since 1.9.5, released on September 22, 2015.
Code execution is possible, but it depends on conditions
According to The Hacker News, F5 warns that code execution is possible on systems where ASLR is disabled or can be defeated. Researchers at Wiz say they got past ASLR and ran code in their tests, and Xint's researchers say a reachable HTTP/2 port was enough. The code runs with the permissions of an NGINX worker process; according to CIRCL, full root access on the host needs another privilege escalation on top. Xint managed that with a separate SCTP bug in the Linux kernel.
HTTP/2 has to be switched on for a server to be exposed, meaning a listen line that includes http2 or the http2 on setting. NGINX Plus is also affected, in 37.x builds before the fixed ones and in R34 through R36. The same report says no fix is listed for R34 to R36.
The updates released that day close other flaws as well. The Hacker News reports that the vendor counts four more of them as potential code execution paths when the same ASLR condition holds, and that all five score 9.2 on CVSS 4.0. Each of those needs its own setup, such as HTTP/3 or a mail proxy.
On exploitation: according to the report, no attacks have been reported so far by the vendor or by the researchers at Wiz and Rapid7, and the flaw was not yet in CISA's catalog of Known Exploited Vulnerabilities that day. A public proof-of-concept exists on GitHub, and its notes limit it to demos on one machine against a Docker container that has just been started. The attack as Rapid7 built it reportedly takes roughly 128 GB of HTTP/2 traffic.
Upgrading is not enough, the process has to be restarted too
As The Hacker News relays it, after an upgrade NGINX has to be fully restarted or moved over with its live binary upgrade. Simply reloading the configuration leaves the old executable running. Updating the package and moving on may therefore not close the hole.
For anyone who cannot upgrade right away, the same report cites CIRCL's suggestion of turning HTTP/2 off as a stopgap. Doing so shuts out CVE-2026-123007 and CVE-2026-123633, while the remaining three flaws stay open. For sites behind a CDN, the report suggests testing the NGINX origin itself, since an HTTP/2 attack cannot reach an origin limited to HTTP/1.1.
Distribution packages were also incomplete that day: the report says Ubuntu and Debian had not yet shipped fixed packages. That may have changed, so check your own distribution's security notice. The official NGINX Docker images already carried 1.31.9 and 1.30.7 tags.
What it means for businesses in Türkiye
This section is our reading, not a fact from the sources. The flaw depends on the software version, not on the country. A business that runs its own VPS or server can check its NGINX version, whether HTTP/2 is on, and whether a full restart followed the upgrade. Those on a hosting provider can ask the provider whether the patch has been applied.
It is also worth watching whether the distribution you use has issued a fixed package. The report points out that a package can carry a fix without changing its version number, and it advises checking the complete package version against the advisory your distribution publishes.
The UNALSOFT view
A website is not only its design. It rests on server software, and who keeps that software current is often unclear. We read this news as an example of why maintenance and update responsibility should be written down once a site is delivered. In our web design projects we settle who owns hosting and server upkeep at the start. The news is a reminder to run a checklist, not a reason to panic.
Do you know which server software your website depends on?
We can map together what your site runs on and who keeps it updated.