[Test] Temporarily skip flaky ACL graph test (#3577)

yiz-liu · web-flow · commit ec1d2b5c0458 · 2025-10-21T17:16:15.000+08:00
### What this PR does / why we need it? Disables `FULL_DECODE_ONLY` end-to-end test that fails intermittently. This prevents CI blockages while the root cause of the flakiness is investigated. ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None needed. - vLLM version: v0.11.0rc3 - vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0 Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
diff --git a/tests/e2e/singlecard/test_aclgraph.py b/tests/e2e/singlecard/test_aclgraph.py
@@ -100,6 +100,8 @@ def test_models_with_aclgraph(
     )
 
 
+@pytest.mark.skip("Skipping this test for now, "
+                  "it fails intermittently and needs investigation.")
 @pytest.mark.parametrize("model", MODELS)
 @pytest.mark.parametrize("max_tokens", [5])
 def test_models_with_aclgraph_full_decode_only(

Original file line number	Diff line number	Diff line change
`@@ -100,6 +100,8 @@ def test_models_with_aclgraph(`
`100`	`100`	`)`
`101`	`101`
`102`	`102`
	`103`	`+@pytest.mark.skip("Skipping this test for now, "`
	`104`	`+ "it fails intermittently and needs investigation.")`
`103`	`105`	`@pytest.mark.parametrize("model", MODELS)`
`104`	`106`	`@pytest.mark.parametrize("max_tokens", [5])`
`105`	`107`	`def test_models_with_aclgraph_full_decode_only(`