{"id":"GHSA-8qpc-6rxx-4pxf","kind":"vulnerability","aliases":["GHSA-8qpc-6rxx-4pxf"],"related":[],"title":"Duplicate Advisory: vllm p2p never connected zmq_exhaustion","description":"## Duplicate Advisory\n\nThis advisory has been withdrawn because it is a duplicate of GHSA-frvg-qw88-hww9. This link is maintained to preserve external references.\n\n## Original Description\nvLLM through 0.29.0 contains a denial of service vulnerability in P2P KV offloading when OffloadingConnector is configured with TieringOffloadingSpec and a peer-to-peer secondary tier. Attackers can supply arbitrary remote host and port values in kv_transfer_params to create unreachable peer sessions that retain ZeroMQ sockets until the context quota is exhausted, causing an uncaught ZMQError that crashes EngineCore and stops all inference.","summary":"A crafted kv_transfer_params request with an arbitrary remote host and port makes vLLM keep ZeroMQ sockets open for unreachable peers until the context quota runs out. The resulting uncaught ZMQError crashes EngineCore and stops all inference.","severity":"high","cvss":7.5,"epss":null,"exploited":false,"exploited_at":null,"products":[{"name":"vllm","fixed":["0.30.0"],"affected":">= 0, < 0.30.0","ecosystem":"pypi"}],"ecosystems":["pypi"],"tags":["ai-tooling","cwe-770","llm-serving","python","self-hosted","vllm"],"not_affected_if":["default-config","no-p2p-offloading"],"how_to_check":"- Run `pip show vllm` and confirm the installed version is below `0.30.0`.\n- Check whether the server is configured with OffloadingConnector and TieringOffloadingSpec with a peer-to-peer secondary tier.\n- unknown for anything beyond that; the sources do not give further checks.","fix":"- Upgrade vllm to `0.30.0` or later.\n- The advisory was withdrawn as a duplicate of GHSA-frvg-qw88-hww9; check that advisory for the authoritative fix guidance.","status":"withdrawn","published_at":"2026-09-22T00:30:55+00:00","first_seen_at":"2026-10-27T14:47:35+00:00","triaged_at":"2026-10-27T14:49:40+00:00","updated_at":"2026-10-27T14:49:40+00:00","category":"ai-tooling","confidence":0.6,"triage":{"at":"2026-10-27T14:49:40Z","model":"claude-haiku-5-5","output":{"fix":"- Upgrade vllm to `0.30.0` or later.\n- The advisory was withdrawn as a duplicate of GHSA-frvg-qw88-hww9; check that advisory for the authoritative fix guidance.","reach":"internet","title":"vLLM: unreachable P2P peer sessions exhaust ZeroMQ sockets and crash inference","impact":"medium","summary":"A crafted kv_transfer_params request with an arbitrary remote host and port makes vLLM keep ZeroMQ sockets open for unreachable peers until the context quota runs out. The resulting uncaught ZMQError crashes EngineCore and stops all inference.","category":"ai-tooling","nicknames":[],"who_cares":["vllm","python","llm-serving","self-hosted"],"confidence":0.6,"needs_auth":"unknown","worst_case":"Inference server crashes and all model requests stop.","analyst_note":"Advisory is withdrawn as duplicate; confirm details and the auth requirement at GHSA-frvg-qw88-hww9.","how_to_check":"- Run `pip show vllm` and confirm the installed version is below `0.30.0`.\n- Check whether the server is configured with OffloadingConnector and TieringOffloadingSpec with a peer-to-peer secondary tier.\n- unknown for anything beyond that; the sources do not give further checks.","needs_analyst":true,"not_affected_if":["default-config","no-p2p-offloading"],"who_is_affected":"Deployments of vLLM through 0.29.0 that use the OffloadingConnector with a TieringOffloadingSpec and a peer-to-peer secondary tier are exposed, because that path accepts remote host and port values from request parameters. Servers without that offloading configuration are not described as affected. The advisory itself is withdrawn as a duplicate, so the live details sit under GHSA-frvg-qw88-hww9.","user_interaction":"no","needs_special_setup":"yes","severity_if_unknown":"high"},"cost_usd":0.000696,"input_tokens":3497,"output_tokens":693,"severity_before":"high"},"analyst_status":null,"analyst_note":null,"search":"'-770':24B '0.29.0':101C '4pxf':13A,18A '6rxx':12A,17A '8qpc':11A,16A 'advisori':2A,72C,74C 'ai':21B 'ai-tool':20B 'arbitrari':41B,129C 'attack':126C 'caus':154C 'configur':115C 'connect':6A 'contain':102C 'context':57B,150C 'craft':34B 'crash':65B,159C 'creat':140C 'cwe':23B 'denial':104C 'descript':98C 'duplic':1A,71C,82C 'enginecor':66B,160C 'exhaust':8A,153C 'extern':95C 'frvg':86C 'ghsa':10A,15A,85C 'ghsa-8qpc-6rxx-4pxf':9A,14A 'ghsa-frvg-qw88-hww9':84C 'host':31B,43B,131C 'hww9':88C 'infer':70B,164C 'keep':48B 'kv':35B,110C,136C 'link':90C 'llm':26B 'llm-serv':25B 'maintain':92C 'make':46B 'never':5A 'offload':111C 'offloadingconnector':113C 'open':51B 'origin':97C 'p2p':4A,109C 'param':37B,138C 'peer':54B,121C,123C,142C 'peer-to-p':120C 'port':45B,133C 'preserv':94C 'python':28B 'quota':58B,151C 'qw88':87C 'refer':96C 'remot':42B,130C 'request':38B 'result':62B 'retain':145C 'run':59B 'secondari':124C 'self':30B 'self-host':29B 'serv':27B 'servic':106C 'session':143C 'socket':50B,147C 'stop':68B,162C 'suppli':128C 'tier':125C 'tieringoffloadingspec':117C 'tool':22B 'transfer':36B,137C 'uncaught':63B,156C 'unreach':53B,141C 'valu':134C 'vllm':3A,19A,32B,47B,99C 'vulner':107C 'withdrawn':77C 'zeromq':49B,146C 'zmq':7A 'zmqerror':64B,157C","nicknames":[],"mitigations":[],"who_affected":"Deployments of vLLM through 0.29.0 that use the OffloadingConnector with a TieringOffloadingSpec and a peer-to-peer secondary tier are exposed, because that path accepts remote host and port values from request parameters. Servers without that offloading configuration are not described as affected. The advisory itself is withdrawn as a duplicate, so the live details sit under GHSA-frvg-qw88-hww9.","tier":"B","verified":{},"analyst_revisit_at":null,"first_public_at":"2026-09-22T00:30:55+00:00","kev":null,"cvss_vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H","cvss_by":"GitHub","poc_count":0,"sources":[{"source":"ghsa","external_id":"GHSA-8qpc-6rxx-4pxf","url":"https://github.com/advisories/GHSA-8qpc-6rxx-4pxf","severity":"high","published_at":"2026-09-22T00:30:55+00:00","modified_at":"2026-10-27T14:45:48+00:00","first_seen_at":"2026-10-27T14:47:35+00:00","fetched_at":"2026-10-27T14:47:35+00:00","note":null}],"documents":[],"notice":"Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai"}