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 7, 02:45 UTC
[CI] Fix startup error test (#36230)
A change in engine startup error messages in #35478 caused this test failure. Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
Mar 6, 2026 at 11:50 UTC
e2090bf3af96843c899d6f5c85d9c12b03b5cabb
2 files changed
+3
-2
pyproject.toml
+1
@@ -172,6 +172,7 @@ fo = "fo"
172
nd = "nd"
173
eles = "eles"
174
datas = "datas"
175
+ure = "ure"
176
177
[tool.uv]
178
no-build-isolation-package = ["torch"]
tests/v1/shutdown/test_startup_error.py
+2
-2
@@ -68,7 +68,7 @@ def test_async_llm_startup_error(
68
)
69
70
# Confirm we get an exception.
71
- with pytest.raises(Exception, match="initialization failed"):
71
+ with pytest.raises(Exception, match=r"initialization fail(ed|ure)"):
72
_ = AsyncLLM.from_engine_args(engine_args)
73
74
# Confirm all the processes are cleaned up.
@@ -111,7 +111,7 @@ def test_llm_startup_error(
111
112
with pytest.raises(
113
Exception,
114
- match="initialization failed"
114
+ match=r"initialization fail(ed|ure)"
115
if enable_multiprocessing
116
else "Simulated Error in startup!",
117
):