[Notice] Interim Mitigation Implemented for the Linux Kernel Vulnerability Infinity Core (CVE-2026-98260)

A vulnerability, “Infinity Core (CVE-2026-98260),” has been reported in the Linux kernel that could lead to local privilege escalation.

To mitigate the impact of this vulnerability, PFCP has temporarily stopped workloads of multiple organizations from sharing the same shared node in SR1-01.

Mitigation

  • On the shared nodes (MN-Server 2) in SR1-01, Pods are no longer scheduled onto a node where Pods of another organization are running.
  • Pods that specify a PriorityClass for shared nodes (shared-standard or shared-best-effort) are no longer allowed to set spec.nodeName.
  • Pods that were running on the shared nodes have been deleted and recreated, so that each organization is placed on its own node.

Completion Time

October 23, 2026, 20:00 JST

Impact on Users

This affects customers using the shared nodes in SR1-01. Customers using only reserved nodes, and customers using clusters other than SR1-01, are not affected.

  • During the work, Pods and Workspace execution environments that were running on the shared nodes were deleted once and recreated.
    • Pods created directly without a higher-level resource such as a Deployment or Job have not been recreated. Please recreate them if needed.
    • Data stored in persistent storage (PVCs, etc.) is not affected. However, data that was not in persistent storage — including data inside Pods and Workspace execution environments — has been lost.
  • Pods are not placed on a node where Pods of another organization are running. If no shared node is available, Pods may remain Pending. If Pods stay Pending, please contact us.
  • Creating a Pod that specifies a PriorityClass for shared nodes together with spec.nodeName is rejected. No other changes to your manifests are required.

Monitoring

We watch for signs of exploitation of this vulnerability and the spread of its impact by combining measures such as monitoring the audit logs of Kubernetes and other systems, monitoring workloads with a runtime security sensor, and monitoring network traffic, and we will take additional measures depending on the situation.

Next Steps

We will continue to put the security of the service first. If we determine, based on threat intelligence and monitoring, that the current interim mitigation is not sufficient, we will promptly take the necessary measures, which may involve reboots. We will also keep the interim mitigation in place until we confirm that the measures have made the service safe.

Once a fixed Linux kernel is released, we will validate it and update the Linux kernel on the Kubernetes nodes.

As with cluster upgrades, the kernel update will reboot compute nodes sequentially (up to 25% at a time) rather than all at once.

  • Compute nodes will be rebooted to apply the kernel update. Applications (Pods) running on Kubernetes and Workspace execution environments will be temporarily stopped, deleted, and recreated.
  • Pods created directly without a higher-level resource such as a Deployment or Job will not be recreated automatically.
  • Data stored in persistent storage (PVCs, etc.) is not affected. However, data not in persistent storage — including data inside Pods and Workspace execution environments — will be lost.
  • Reserved nodes allocated to users will be temporarily unavailable while they are being rebooted.
    • If you have fewer than 4 reserved nodes, they will be rebooted one at a time. If you have 4 or more, up to 25% (rounded down) will be temporarily unavailable at a time.
    • If you have only 1 reserved node, it will be temporarily unavailable.
  • Up to 25% (rounded down) of shared nodes will be temporarily unavailable.

Update (November 11, 2026)

The update to the fixed Linux kernel will be performed as emergency maintenance on Monday, November 16, 2026. For details, see Maintenance Notice (November 16, 2026).

References


PFCP is committed to providing timely security updates so that our users can continue to use our services with confidence.