CVE-2026-94627

vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collision

Published Sep 21, 2026 CVSS 4.0 · 8.7 High EPSS < 1% KEV · Not listed
Description
vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing.

Analysis

AI-generated

Technical impact

vLLM’s Mooncake connector, up to version 0.29.0, fails to properly manage ownership of GPU KV cache blocks when multiple completion requests share a single transfer ID during prefill/decode disaggregated deployments. The flaw allows an attacker to trigger an accumulation of orphaned KV cache blocks, exhausting GPU memory until the inference process must restart, thereby preventing legitimate inference requests from being served. This is a classic memory‑leak issue (CWE‑401) that leads to a denial‑of‑service outcome. The description indicates that the vulnerability arises from improper resource cleanup (CWE‑772) during concurrent child requests.

Affected Systems

The affected product is the official vllm project, versions 0.29.0 and earlier. Any deployment that uses the Mooncake connector for GPU KV cache transfers is vulnerable. No specific operating system or GPU vendor is mandated by the description, so any GPU‑enabled system running these versions of vllm with the connector will be impacted.

Risk and Exploitability

The CVSS score of 8.7 indicates high severity. The EPSS score of < 1% signals a very low but nonzero likelihood of exploitation at the time of analysis. The vulnerability is not yet listed in CISA’s KEV catalog. Attackers can trigger the denial‑of‑service path by sending legitimate completion requests to the exposed vLLM API. Based on the description, it is inferred that authentication is not explicitly required; the flaw can be triggered with normal clients possessing API access, implying a less restrictive attack vector. The risk is consequential for production environments with shared GPU resources, as the exhaustion can cause repeated restarts and prolonged outages.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade to the latest stable vllm release once the issue is fixed; verify that the Mooncake connector’s memory management bug has been addressed.
  • Configure the inference endpoint to require authenticated API keys or tokens so that only trusted clients can issue completion requests, reducing the available attack surface.
  • Enforce that each child request uses a unique transfer ID or establish a queuing system that processes requests sequentially to prevent ID collisions that orphan cache blocks.
  • Implement automated GPU memory monitoring; configure the service to restart automatically at a safe threshold to recover from accidental exhaustion.

Generated by OpenCVE AI on September 23, 2026 at 04:05 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'none', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Type Values Removed Values Added
Description vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing.
Title vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collision
First Time appeared Vllm
Vllm vllm
Weaknesses CWE-401
CPEs cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
Vendors & Products Vllm
Vllm vllm
References
Metrics cvssV3_1

{'score': 7.5, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H'}

cvssV4_0

{'score': 8.7, 'vector': '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'}

Weaknesses
  • CWE-401

    Missing Release of Memory after Effective Lifetime

  • CWE-772

    Missing Release of Resource after Effective Lifetime

Sources

Status
PUBLISHED
Assigner
VulnCheck
Published
Updated
Sep 22, 2026, 12:51 PM
Reserved
Sep 21, 2026, 9:42 PM
Status
Analyzed
Published
Sep 21, 2026, 10:17 PM
Modified
Sep 29, 2026, 12:46 PM
Severity
Important
Public date
Sep 21, 2026, 10:04 PM