CloudLinux says every affected version now has a fixed kernel or a KernelCare livepatch, and lists the CloudLinux 9 livepatch for kernel 5.14.0-687.46.1.el9_8. The watched kernel 5.14.0-611.55.1.el9_7 is not named, so it should be treated as unpatched until confirmed.
A kernel flaw (NoisyTCP, CVE-2026-99288) can crash a Linux server through its network port without a login. CloudLinux now lists a fix for CloudLinux 9, and BIS is checking which kernel your server runs and whether it needs a reboot. We will send the next update when we confirm your server is patched.
CloudLinux says: “Every affected CloudLinux version now has a fixed kernel or a KernelCare livepatch.” Source
What the IT community is saying
Posts from about Nov 15–17 mostly relay the disclosure: a pre-auth remote kernel crash via TCP that the researcher says can reach default Nginx, Apache and SSH setups, with patches in upstream 6.12.117, 6.18.63 and 7.3.4 and mitigations of disabling SYN cookies or using a single RX queue. Admins are urging prompt patching, and one reply warns that distro kernels backport fixes under older version numbers. I found no CloudLinux-specific post, and none confirming CloudLinux 9 is unpatched or fixed.
From public posts by IT administrators; not confirmed by the vendor.
Questions? Contact BIS Computer Solutions, Inc.
Phone: +1-818-248-5023 · Fax: +1-818-249-1169 · Email: support@biscomputer.com · Web: www.biscomputer.com
Technical details
Reference numbers
| ID | Severity | Exploited |
|---|---|---|
| CVE-2026-99288 | CVSS 7.5 · high | No |
| DSA-6628-1 | No | |
| DLA-4906-1 | No | |
| CloudLinux status incident NoisyTCP | No |
Affected versions and fixes
| Product / distribution | Affected | Fixed in | Status | Source |
|---|---|---|---|---|
| CloudLinux 9 kernel 5.14.0-611.55.1.el9_7 (watched) | 5.14.0-611.55.1.el9_7 | Unconfirmed | CloudLinux status | |
| CloudLinux 9 main feed kernel | before 5.14.0-687.46.1.el9_8 | 5.14.0-687.46.1.el9_8 (KernelCare livepatch) | Fixed | CloudLinux status |
| CloudLinux 10 main feed kernel | before 6.12.0-211.53.1.el10_2 | 6.12.0-211.53.1.el10_2 (KernelCare livepatch) | Fixed | CloudLinux status |
| AlmaLinux 9 kernel | unconfirmed | Unconfirmed | AlmaLinux blog, 2026-11-16 | |
| Red Hat Enterprise Linux 8, 9, 10 | all kernels before the fix | Waiting on vendor | TuxCare (Red Hat status as of Nov 16), 2026-11-16 | |
| Debian 12 linux | before 6.1.192-1 | 6.1.192-1 (DLA-4906-1) | Fixed | Debian security tracker |
| Debian 13 linux | before 6.12.117-1 | 6.12.117-1 (DSA-6628-1) | Fixed | Debian security tracker |
| Upstream Linux stable | before 6.12.117, 6.18.63, 7.3.4 | 6.12.117, 6.18.63, 7.3.4 | Fixed | Linux kernel CVE record, 2026-11-10 |
Workarounds until a fix is installed
- Set net.ipv4.tcp_syncookies=0 (per network namespace; weakens SYN-flood protection).
- Reduce the interface to one RX queue and disable RPS and RFS.
How to check if a system is affected
uname -rrpm -q kernelkcarectl --patch-info | grep CVE-2026-99288
Root cause
A race in the lockless established-socket lookup lets a recycled socket be used before it is hashed, causing a NULL pointer dereference in tcp_v4_rcv() and a panic.
Exploitation
No exploitation reported per CloudLinux (Nov 17) and Red Hat; a limited lab exploit for VMs is described by the researcher, and sources disagree on whether a crash PoC is public.
Latest updates
- Nov 21, 2026, 9:51 AM PT: BIS staff report a CloudLinux 9 status update still showing no livepatch, but the CloudLinux status page now lists a KernelCare livepatch for CloudLinux 9 on kernel 5.14.0-687.46.1.el9_8; the watched 9_7 kernel is not named, so it is still unconfirmed.
What the IT community is saying
Posts from about Nov 15–17 mostly relay the disclosure: a pre-auth remote kernel crash via TCP that the researcher says can reach default Nginx, Apache and SSH setups, with patches in upstream 6.12.117, 6.18.63 and 7.3.4 and mitigations of disabling SYN cookies or using a single RX queue. Admins are urging prompt patching, and one reply warns that distro kernels backport fixes under older version numbers. I found no CloudLinux-specific post, and none confirming CloudLinux 9 is unpatched or fixed.
From public posts by IT administrators; not confirmed by the vendor.
What you need to do
Reboot into the fixed kernel once it is installed; BIS will confirm on managed machines.
Is this a hack?
No It is a remote crash (denial of service), not data theft; no exploitation has been reported.
Root cause
A race in the kernel's lockless established-TCP socket lookup can lead to a NULL pointer dereference and a kernel panic. Red Hat rates it Important, CVSS 7.5 (preliminary).
