@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 003f8ee128
Fork 0

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."""