A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 19:30 UTC
[BugFix] Use IP4 localhost form for zmq bind (#7163)
Nick Hill committed
Aug 5, 2024 at 08:41 UTC
003f8ee1287f90a7e8aa9b9e7d6246ac74ebefbe
1 file changed
+3
-1
vllm/entrypoints/openai/rpc/server.py
+3
-1
@@ -30,7 +30,9 @@ class AsyncEngineRPCServer:
30
31
# Init socket for readiness state.
32
self.socket = self.context.socket(zmq.constants.ROUTER)
33
- self.socket.bind(f"tcp://localhost:{port}")
33
+ # Note numeric form of localhost should be used for zmq bind(),
34
+ # see https://stackoverflow.com/a/8958414
35
+ self.socket.bind(f"tcp://127.0.0.1:{port}")
36
37
def cleanup(self):
38
"""Cleanup all resources."""