Fix pylint hang due to jobs concurrency

Makefile

@@ -21,23 +21,23 @@ install:
 
 .PHONY: lint
 lint: check_pylint_installed check_pytest_installed
-	@pylint --rcfile=.pylintrc -s n  --jobs=128 ./llumnix
-	
+	@pylint --rcfile=.pylintrc -s n --jobs=128 ./llumnix
+
 	@pylint --rcfile=.pylintrc \
 			--disable=protected-access,super-init-not-called,unused-argument,redefined-outer-name,invalid-name \
 			-s n --jobs=128 ./tests
 
 .PHONY: test
 test: check_pytest_installed
-	@pytest -v --ignore=third_party/ --ignore=tests/e2e_test --disable-warnings
+	@pytest -v --ignore=third_party --ignore=tests/e2e_test --disable-warnings
 	@python examlpes/offline_inference.py
 	@pytest -v -x -s --tb=long ./tests/e2e_test/test_e2e.py
 	@pytest -v -x -s --tb=long ./tests/e2e_test/test_bench.py
 	@pytest -v -x -s --tb=long ./tests/e2e_test/test_migration.py
 
 .PHONY: unit_test
 unit_test: check_pytest_installed
-	@pytest -v --ignore=third_party/ --ignore=tests/e2e_test --disable-warnings
+	@pytest -v --ignore=third_party --ignore=tests/e2e_test --disable-warnings
 
 .PHONY: offline_test
 offline_test:

llumnix/backends/utils.py

@@ -40,10 +40,7 @@ def initialize_placement_group(
     """Initialize the distributed cluster probably with Ray.
 
     Args:
-        parallel_config: The configurations for parallel execution.
-        engine_use_ray: Whether to use Ray for async engine.
-        ray_address: The address of the Ray cluster. If None, uses
-            the default Ray cluster address.
+        world_size: The number of workers in Llumlet.
 
     Returns:
         A tuple of (`distributed_init_method`, `placement_group`). The

tests/unit_test/entrypoints/vllm/api_server_manager_service.py

@@ -12,6 +12,7 @@
 # limitations under the License.
 
 import argparse
+import time
 import uvicorn
 import ray
 from ray.util.queue import Queue as RayQueue
@@ -32,7 +33,6 @@
 engine_manager = None
 MANAGER_ACTOR_NAME = llumnix.llm_engine_manager.MANAGER_ACTOR_NAME
 ENTRYPOINTS_ACTOR_NAME = "entrypoints"
-request_output_queue = RayQueue()
 
 
 @ray.remote(num_cpus=0, lifetime="detached")
@@ -110,8 +110,11 @@ def stats() -> Response:
     parser = EngineManagerArgs.add_cli_args(parser)
     args = parser.parse_args()
 
+    # magic actor, without this actor, FastAPIServer cannot initialize correctly.
+    # If this actor is placed globally, pylint will hangs if testing api_server_manager and api_server_service concurrently (--jobs > 1).
+    request_output_queue = RayQueue()
+
     request_output_queue_type = QueueType(args.request_output_queue_type)
     engine_manager = init_manager_service(request_output_queue_type, args)
-
-    import time
-    time.sleep(5)
+
+    time.sleep(2)