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 12, 02:15 UTC
[Core] Fix standalone runs of test_reset_prefix_cache_e2e (#29899)
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Mark McLoughlin committed
Dec 4, 2025 at 08:22 UTC
899e2ef558e7345b99bc0d53c2e1c60ffdca7470
1 file changed
+3
-1
tests/v1/core/test_reset_prefix_cache_e2e.py
+3
-1
@@ -11,7 +11,9 @@ PROMPTS = [
11
]
12
13
14
-def test_reset_prefix_cache_e2e():
14
+def test_reset_prefix_cache_e2e(monkeypatch):
15
+ # "spawn" is required for test to be deterministic
16
+ monkeypatch.setenv("VLLM_WORKER_MULTIPROC_METHOD", "spawn")
17
engine_args = EngineArgs(
18
model="Qwen/Qwen3-0.6B",
19
gpu_memory_utilization=0.2,