Live Exposure Defense: From CVE to Confirmed Exposure in 12 Hours – See more

Back
New CVE Detected

CVE-2026-94624 – Denial of Service via Unbounded P2P KV-Offload Sessions – vLLM ≤ 0.29.0

Be the first to know when new zero-days emerge:

Subscribe to Email Alerts
Subscribe to RSS Feed

Summary

CVE-2026-94624 is a denial-of-service vulnerability in vLLM’s peer-to-peer (P2P) KV-cache offloading feature. When a vLLM deployment uses the OffloadingConnector with a TieringOffloadingSpec and a peer-to-peer secondary tier, an unauthenticated remote attacker can submit crafted kv_transfer_params values to exhaust the engine’s ZeroMQ socket resources, crashing the inference engine (EngineCore) and halting all inference for the deployment. The issue affects vLLM through version 0.29.0 and carries a CVSS v4.0 base score of 8.7 (High).

Technical details

  • Root cause: vLLM’s P2P KV-offload session manager creates a new peer session — and 3 associated ZeroMQ sockets (DEALER, monitor, and PAIR) — for every distinct remote host/port pair supplied in kv_transfer_params, without capping the number of sessions or reaping unreachable/idle ones.
  • Trigger condition: An attacker submits inference requests containing kv_transfer_params with arbitrary, non-listening remote host and port values, causing the server to open sockets for peers that never complete a handshake.
  • Attack vector: Network-based, unauthenticated, no user interaction required (AV:N/AC:L/PR:N/UI:N). No special privileges are needed — any client able to reach the OpenAI-compatible inference API can trigger the condition.
  • Impact: Each unreachable peer consumes ZMQ sockets that are never released. Roughly 340 such peer sessions exhaust the ZMQ context’s MAX_SOCKETS limit (1023), producing an uncaught ZMQError that crashes EngineCore and stops all inference processing for the deployment — a full availability impact with no confidentiality or integrity loss (CWE-770: Allocation of Resources Without Limits or Throttling).

Affected software

  • vLLM (vllm-project/vllm), all versions through 0.29.0, when configured with OffloadingConnector using a TieringOffloadingSpec and a peer-to-peer (P2P) secondary offload tier.

Severity

  • CVSS v4.0: 8.7 (High) — CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N
  • CVSS v3.1: 7.5 (High) — CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Mitigation and recommended actions

  • Immediate: Upgrade to a vLLM release that incorporates the fix tracked in GitHub Security Advisory GHSA-frvg-qw88-hww9 (patch developed in pull request #51504), which introduces a session capacity limit, a handshake deadline, idle-session eviction, and an error boundary around peer-session creation. Deployments should track vLLM release notes and apply the first version after 0.29.0 that includes this fix.
  • If no patch is available yet:
    • Disable or restrict use of the P2P KV-offload tiering feature (OffloadingConnector with TieringOffloadingSpec in P2P mode) on internet-facing or otherwise untrusted-client deployments until patched.
    • Restrict network access to the vLLM inference API to trusted clients only (e.g., via network segmentation, authentication proxy, or allow-listing) so that untrusted parties cannot submit arbitrary kv_transfer_params.
    • Monitor EngineCore process health and ZeroMQ socket counts, and configure automatic restart/alerting on EngineCore crashes to reduce downtime from exploitation attempts.

References

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

  • Mapping of all Assets with this Technology
  • Identification of potentially exposed assets to this CVE
  • Confirmation of verified exploitable assets
Get Exposure Report

How IONIX’s External Exposure Management Platform Detects and Validates
Zero-Days to Shrink MTTR

1

Map your entire attack surface (continously)

IONIX uses multi-factor discovery methods, including DNS analysis, certificate mapping, metadata inspection, and more, to automatically map every internet-facing asset across your environment. This includes cloud instances, third-party platforms, shadow IT, and even forgotten infrastructure that traditional tools miss.

2

Monitor for new CVEs

Dozens of threat intel feeds using agentic technology are continuously analyzed to detect the appearance of proof-of-concept code, exploit kits, and indicators of active targeting. IONIX goes further by applying AI to proactively evaluate whether emerging vulnerabilities are likely to be exploited, even before PoCs go public.

3

Identify Potential External Exposures

Not all CVEs matter. IONIX filters vulnerabilities by asking attacker-centric questions: Can it be reached from the internet? Does it require authentication? Is it being exploited in the wild? This dramatically reduces noise and focuses teams on threats that can actually be weaponized.

4

Create Safe, Scalable Exploit Validations

IONIX transforms real-world PoCs into safe, non-intrusive test payloads that can be run in production environments without disruption. These simulations are precisely targeted to the systems that are vulnerable, ensuring rapid validation without unnecessary load.

5

Execute Exploit Validations

By combining context about software stack, versioning, exposure status, and reachability, IONIX ensures that only the right payloads are executed against the right assets, maximizing efficiency and minimizing risk.

6

Drive Fast and Actionable Remediation

Results are routed through integrations with ticketing, SOAR, and SIEM tools. Issues are written in plain language, bundled into remediation clusters, and prioritized based on asset criticality, exploitability, and blast radius. This shortens mean time to remediation (MTTR) and empowers teams to act with confidence.

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

  • Mapping of all Assets with this Technology
  • Identification of potentially exposed assets to this CVE
  • Confirmation of verified exploitable assets
Get Exposure Report
Close

Subscribe to Threat Center RSS

Copy/paste the link below into your preferred RSS reader or follow these instructions to subscribe to Slack alerts.

Open RSS feedSlack setup instructions

Close

Get Real-Time CVE Alerts to Your Email

Be the first to know when new zero-days emerge