[Core] Remove unused credis related code. (#10849)

* Done.

* Lint.

python/ray/tests/test_actor_advanced.py

@@ -189,9 +189,6 @@ def test_exception_raised_when_actor_node_dies(ray_start_cluster_head):
                 ray.get(x_id)
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_actor_init_fails(ray_start_cluster_head):
     cluster = ray_start_cluster_head
     remote_node = cluster.add_node()
@@ -347,9 +344,6 @@ def test_distributed_handle(ray_start_cluster_2_nodes):
 
 
 @pytest.mark.skip("This test does not work yet.")
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_remote_checkpoint_distributed_handle(ray_start_cluster_2_nodes):
     cluster = ray_start_cluster_2_nodes
     counter, ids = setup_counter_actor(test_checkpoint=True)

python/ray/tests/test_actor_resources.py

@@ -79,9 +79,6 @@ def test_actor_class_methods(ray_start_regular):
     assert ray.get(a.g.remote(2)) == 4
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Failing with new GCS API on Linux.")
 @pytest.mark.skipif(sys.platform == "win32", reason="Failing on Windows.")
 def test_actor_gpus(ray_start_cluster):
     cluster = ray_start_cluster

python/ray/tests/test_advanced_3.py

@@ -257,9 +257,6 @@ def test_not_logging_to_driver(shutdown_only):
     assert len(err_lines) == 0
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="New GCS API doesn't have a Python API yet.")
 def test_workers(shutdown_only):
     num_workers = 3
     ray.init(num_cpus=num_workers)

python/ray/tests/test_component_failures.py

@@ -14,9 +14,6 @@ SIGKILL = signal.SIGKILL if sys.platform != "win32" else signal.SIGTERM
 
 # This test checks that when a worker dies in the middle of a get, the plasma
 # store and raylet will not die.
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Not working with new GCS API.")
 def test_dying_worker_get(ray_start_2_cpus):
     @ray.remote
     def sleep_forever(signal):
@@ -65,9 +62,6 @@ def test_dying_worker_get(ray_start_2_cpus):
 
 # This test checks that when a driver dies in the middle of a get, the plasma
 # store and raylet will not die.
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Not working with new GCS API.")
 def test_dying_driver_get(ray_start_regular):
     # Start the Ray processes.
     address_info = ray_start_regular
@@ -109,9 +103,6 @@ ray.get(ray.ObjectRef(ray.utils.hex_to_binary("{}")))
 
 # This test checks that when a worker dies in the middle of a wait, the plasma
 # store and raylet will not die.
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Not working with new GCS API.")
 def test_dying_worker_wait(ray_start_2_cpus):
     @ray.remote
     def sleep_forever():
@@ -150,9 +141,6 @@ def test_dying_worker_wait(ray_start_2_cpus):
 
 # This test checks that when a driver dies in the middle of a wait, the plasma
 # store and raylet will not die.
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Not working with new GCS API.")
 def test_dying_driver_wait(ray_start_regular):
     # Start the Ray processes.
     address_info = ray_start_regular
@@ -193,5 +181,4 @@ ray.wait([ray.ObjectRef(ray.utils.hex_to_binary("{}"))])
 
 
 if __name__ == "__main__":
-    import pytest
     sys.exit(pytest.main(["-v", __file__]))

python/ray/tests/test_component_failures_3.py

@@ -1,4 +1,3 @@
-import os
 import sys
 import time
 
@@ -75,9 +74,6 @@ def test_actor_creation_node_failure(ray_start_cluster):
         cluster.remove_node(get_other_nodes(cluster, True)[-1])
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_driver_lives_sequential(ray_start_regular):
     ray.worker._global_node.kill_raylet()
     ray.worker._global_node.kill_plasma_store()
@@ -88,9 +84,6 @@ def test_driver_lives_sequential(ray_start_regular):
     # If the driver can reach the tearDown method, then it is still alive.
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_driver_lives_parallel(ray_start_regular):
     all_processes = ray.worker._global_node.all_processes
 

python/ray/tests/test_multi_node.py

@@ -419,19 +419,18 @@ def test_calling_start_ray_head(call_ray_stop_only):
     check_call_ray(["start", "--head", "--redis-max-clients", "100"])
     check_call_ray(["stop"])
 
-    if "RAY_USE_NEW_GCS" not in os.environ:
-        # Test starting Ray with redis shard ports specified.
-        check_call_ray(
-            ["start", "--head", "--redis-shard-ports", "6380,6381,6382"])
-        check_call_ray(["stop"])
+    # Test starting Ray with redis shard ports specified.
+    check_call_ray(
+        ["start", "--head", "--redis-shard-ports", "6380,6381,6382"])
+    check_call_ray(["stop"])
 
-        # Test starting Ray with all arguments specified.
-        check_call_ray([
-            "start", "--head", "--redis-shard-ports", "6380,6381,6382",
-            "--object-manager-port", "12345", "--num-cpus", "2", "--num-gpus",
-            "0", "--redis-max-clients", "100", "--resources", "{\"Custom\": 1}"
-        ])
-        check_call_ray(["stop"])
+    # Test starting Ray with all arguments specified.
+    check_call_ray([
+        "start", "--head", "--redis-shard-ports", "6380,6381,6382",
+        "--object-manager-port", "12345", "--num-cpus", "2", "--num-gpus", "0",
+        "--redis-max-clients", "100", "--resources", "{\"Custom\": 1}"
+    ])
+    check_call_ray(["stop"])
 
     # Test starting Ray with invalid arguments.
     with pytest.raises(subprocess.CalledProcessError):

python/ray/tests/test_multinode_failures_2.py

@@ -1,4 +1,3 @@
-import os
 import sys
 import time
 
@@ -124,9 +123,6 @@ def test_actor_creation_node_failure(ray_start_cluster):
         cluster.remove_node(get_other_nodes(cluster, True)[-1])
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_driver_lives_sequential(ray_start_regular):
     ray.worker._global_node.kill_raylet()
     ray.worker._global_node.kill_plasma_store()
@@ -137,9 +133,6 @@ def test_driver_lives_sequential(ray_start_regular):
     # If the driver can reach the tearDown method, then it is still alive.
 
 
-@pytest.mark.skipif(
-    os.environ.get("RAY_USE_NEW_GCS") == "on",
-    reason="Hanging with new GCS API.")
 def test_driver_lives_parallel(ray_start_regular):
     all_processes = ray.worker._global_node.all_processes
 

python/ray/tests/test_ray_init.py

@@ -15,9 +15,6 @@ def password():
 
 
 class TestRedisPassword:
-    @pytest.mark.skipif(
-        os.environ.get("RAY_USE_NEW_GCS") == "on",
-        reason="New GCS API doesn't support Redis authentication yet.")
     def test_redis_password(self, password, shutdown_only):
         @ray.remote
         def f():
@@ -42,9 +39,6 @@ class TestRedisPassword:
             host=redis_ip, port=redis_port, password=password)
         assert redis_client.ping()
 
-    @pytest.mark.skipif(
-        os.environ.get("RAY_USE_NEW_GCS") == "on",
-        reason="New GCS API doesn't support Redis authentication yet.")
     def test_redis_password_cluster(self, password, shutdown_only):
         @ray.remote
         def f():