{"id":"CVE-2026-94627","published":"2026-09-21T22:17:01.740","lastModified":"2026-09-22T20:25:55.870","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.","cvssScore":7.5,"cvssSeverity":"HIGH","cvssVector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H","cwes":["CWE-401"],"vendors":[],"products":[],"references":[{"url":"https://github.com/vllm-project/vllm","tags":[]},{"url":"https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989","tags":[]},{"url":"https://github.com/vllm-project/vllm/pull/49796","tags":[]},{"url":"https://www.vulncheck.com/advisories/vllm-through-0.29.0-gpu-kv-cache-leak-via-mooncake-transfer-id-collision","tags":[]}],"exploitRefs":[{"url":"https://github.com/vllm-project/vllm","tags":[]},{"url":"https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989","tags":[]},{"url":"https://github.com/vllm-project/vllm/pull/49796","tags":[]}],"hasPoc":true,"ai":{"summary":"The vLLM Mooncake connector has a memory management flaw where GPU KV cache blocks aren't properly released when concurrent requests share the same transfer ID, leading to gradual memory exhaustion that eventually prevents legitimate requests from processing.","exploitability":"Exploitation is relatively straightforward requiring only submission of multiple completion requests with shared transfer IDs, though it may require specific deployment configurations with prefill/decode disaggregated setups.","blast_radius":"In production deployments with high request volumes, this flaw could cause complete service unavailability as GPU memory fills up and legitimate requests fail to execute, potentially affecting all users of the affected LLM inference service.","remediation":"Upgrade to vLLM Mooncake connector version 0.29.1 or later which contains the fix for the KV cache block ownership management issue.","detection":"Monitor GPU memory usage patterns showing gradual increase over time with no corresponding release of memory blocks, and watch for process restarts or memory exhaustion errors in system logs.","tags":["memory-leak","llm","gpu","inference"],"model":"garnet-coder","analyzedAt":"2026-10-18T16:15:28.290Z"}}