[Attention] Update tests to remove deprecated env vars (#30563)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>

tests/distributed/test_context_parallel.py

@@ -219,14 +219,12 @@ def _test_cp_gsm8k(
         ]
     )
 
-    server_env = {}
     if attn_backend:
-        server_env["VLLM_ATTENTION_BACKEND"] = attn_backend
+        server_args.append(f"--attention-backend={attn_backend}")
 
     with RemoteOpenAIServer(
         model_id,
         server_args,
-        env_dict=server_env,
         max_wait_seconds=720,
     ) as remote_server:
         host = f"http://{remote_server.host}"

tests/distributed/test_pp_cudagraph.py

@@ -20,23 +20,21 @@ from ..utils import compare_two_settings, create_new_process_for_each_test
 )
 @create_new_process_for_each_test()
 def test_pp_cudagraph(
-    monkeypatch: pytest.MonkeyPatch,
     PP_SIZE: int,
     MODEL_NAME: str,
     ATTN_BACKEND: LiteralString,
 ):
-    with monkeypatch.context() as m:
-        cudagraph_args = [
-            # use half precision for speed and memory savings in CI environment
-            "--dtype",
-            "float16",
-            "--pipeline-parallel-size",
-            str(PP_SIZE),
-            "--distributed-executor-backend",
-            "mp",
-        ]
-        m.setenv("VLLM_ATTENTION_BACKEND", ATTN_BACKEND)
+    cudagraph_args = [
+        # use half precision for speed and memory savings in CI environment
+        "--dtype",
+        "float16",
+        "--pipeline-parallel-size",
+        str(PP_SIZE),
+        "--distributed-executor-backend",
+        "mp",
+        f"--attention-backend={ATTN_BACKEND}",
+    ]
 
-        eager_args = cudagraph_args + ["--enforce-eager"]
+    eager_args = cudagraph_args + ["--enforce-eager"]
 
-        compare_two_settings(MODEL_NAME, eager_args, cudagraph_args)
+    compare_two_settings(MODEL_NAME, eager_args, cudagraph_args)