[Fast Start] Use GPU uuid as socket folder identifier (#56669)
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
This commit is contained in:
@@ -7,6 +7,7 @@ warm restarts (weights mapped from the daemon via CUDA IPC) must both serve
|
||||
identical outputs.
|
||||
"""
|
||||
|
||||
import glob
|
||||
import os
|
||||
import shutil
|
||||
import subprocess
|
||||
@@ -21,7 +22,6 @@ import pytest
|
||||
|
||||
from vllm import SamplingParams
|
||||
from vllm.assets.image import ImageAsset
|
||||
from vllm.model_executor.model_loader.weight_cache.protocol import get_socket_path
|
||||
from vllm.platforms import current_platform
|
||||
|
||||
DAEMON_TIMEOUT_S = 600
|
||||
@@ -79,9 +79,7 @@ class WeightCacheDaemon:
|
||||
# Poll for the socket files rather than a log line: model loading can
|
||||
# pull in JIT compilers that swap the process's stderr and swallow
|
||||
# everything logged afterwards, making log-based readiness flaky.
|
||||
expected = [
|
||||
get_socket_path(gpu_id, self.socket_dir) for gpu_id in range(self.tp_size)
|
||||
]
|
||||
pattern = os.path.join(self.socket_dir, "vllm_weight_cache_*.sock")
|
||||
deadline = time.monotonic() + timeout_s
|
||||
while time.monotonic() < deadline:
|
||||
if self._proc.poll() is not None:
|
||||
@@ -89,7 +87,7 @@ class WeightCacheDaemon:
|
||||
f"Weight cache daemon exited with {self._proc.returncode}:\n"
|
||||
f"{self._logs()}"
|
||||
)
|
||||
if all(os.path.exists(path) for path in expected):
|
||||
if len(glob.glob(pattern)) >= self.tp_size:
|
||||
return
|
||||
time.sleep(1.0)
|
||||
raise TimeoutError(
|
||||
|
||||
Reference in New Issue
Block a user