tests/v1/e2e/spec_decode: assert async scheduling is used (#39206)

Signed-off-by: Rishi Puri <riship@nvidia.com>
Signed-off-by: Rishi Puri <puririshi98@berkeley.edu>
Signed-off-by: sfeng33 <4florafeng@gmail.com>
Co-authored-by: Benjamin Chislett <chislett.ben@gmail.com>
Co-authored-by: Flora Feng <4florafeng@gmail.com>

tests/v1/e2e/spec_decode/test_async_spec_decode.py

@@ -116,6 +116,11 @@ def test_no_sync_with_spec_decode(
         async_scheduling=True,
     )
 
+    # Assert async scheduling is actually active before running inference.
+    assert llm.llm_engine.vllm_config.scheduler_config.async_scheduling, (
+        f"Expected async_scheduling=True for spec decode, got False. method={method}"
+    )
+
     outputs = llm.generate(
         ["Hello, my name is"],
         SamplingParams(temperature=0, max_tokens=10),