Change training tasks to xray for Jenkins tests (#2567)

python/ray/rllib/env/atari_wrappers.py

@@ -153,7 +153,10 @@ class WarpFrame(gym.ObservationWrapper):
         self.width = dim  # in rllib we use 80
         self.height = dim
         self.observation_space = spaces.Box(
-            low=0, high=255, shape=(self.height, self.width, 1))
+            low=0,
+            high=255,
+            shape=(self.height, self.width, 1),
+            dtype=np.float32)
 
     def observation(self, frame):
         frame = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
@@ -170,7 +173,10 @@ class FrameStack(gym.Wrapper):
         self.frames = deque([], maxlen=k)
         shp = env.observation_space.shape
         self.observation_space = spaces.Box(
-            low=0, high=255, shape=(shp[0], shp[1], shp[2] * k))
+            low=0,
+            high=255,
+            shape=(shp[0], shp[1], shp[2] * k),
+            dtype=np.float32)
 
     def reset(self):
         ob = self.env.reset()

python/ray/rllib/examples/serving/cartpole_server.py

@@ -10,6 +10,7 @@ To try this out, in two separate shells run:
 
 import os
 from gym import spaces
+import numpy as np
 
 import ray
 from ray.rllib.agents.dqn import DQNAgent
@@ -25,8 +26,9 @@ CHECKPOINT_FILE = "last_checkpoint.out"
 
 class CartpoleServing(ServingEnv):
     def __init__(self):
-        ServingEnv.__init__(self, spaces.Discrete(2),
-                            spaces.Box(low=-10, high=10, shape=(4, )))
+        ServingEnv.__init__(
+            self, spaces.Discrete(2),
+            spaces.Box(low=-10, high=10, shape=(4, ), dtype=np.float32))
 
     def run(self):
         print("Starting policy server at {}:{}".format(SERVER_ADDRESS,

python/ray/rllib/utils/actors.py

@@ -36,7 +36,11 @@ class TaskPool(object):
 
         for worker, obj_id in self.completed():
             plasma_id = ray.pyarrow.plasma.ObjectID(obj_id.id())
-            ray.worker.global_worker.plasma_client.fetch([plasma_id])
+            if not ray.global_state.use_raylet:
+                ray.worker.global_worker.plasma_client.fetch([plasma_id])
+            else:
+                (ray.worker.global_worker.local_scheduler_client.
+                 reconstruct_objects([obj_id], True))
             self._fetching.append((worker, obj_id))
 
         remaining = []

python/ray/rllib/utils/policy_server.py

@@ -28,7 +28,11 @@ class PolicyServer(ThreadingMixIn, HTTPServer):
                def __init__(self):
                    ServingEnv.__init__(
                        self, spaces.Discrete(2),
-                       spaces.Box(low=-10, high=10, shape=(4,)))
+                       spaces.Box(
+                           low=-10,
+                           high=10,
+                           shape=(4,),
+                           dtype=np.float32))
                def run(self):
                    server = PolicyServer(self, "localhost", 8900)
                    server.serve_forever()

test/jenkins_tests/run_multi_node_tests.sh

@@ -11,208 +11,208 @@ ROOT_DIR=$(cd "$(dirname "${BASH_SOURCE:-$0}")"; pwd)
 DOCKER_SHA=$($ROOT_DIR/../../build-docker.sh --output-sha --no-cache)
 echo "Using Docker image" $DOCKER_SHA
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env PongDeterministic-v0 \
     --run A3C \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 16}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"kl_coeff": 1.0, "num_sgd_iter": 10, "sgd_stepsize": 1e-4, "sgd_batchsize": 64, "timesteps_per_batch": 2000, "num_workers": 1, "model": {"free_log_std": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"simple_optimizer": false, "num_sgd_iter": 2, "model": {"use_lstm": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"simple_optimizer": true, "num_sgd_iter": 2, "model": {"use_lstm": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"kl_coeff": 1.0, "num_sgd_iter": 10, "sgd_stepsize": 1e-4, "sgd_batchsize": 64, "timesteps_per_batch": 2000, "num_workers": 1, "use_gae": false}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env Pendulum-v0 \
     --run ES \
     --stop '{"training_iteration": 2}' \
     --config '{"stepsize": 0.01, "episodes_per_batch": 20, "timesteps_per_batch": 100}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env Pong-v0 \
     --run ES \
     --stop '{"training_iteration": 2}' \
     --config '{"stepsize": 0.01, "episodes_per_batch": 20, "timesteps_per_batch": 100}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run A3C \
     --stop '{"training_iteration": 2}' \
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run DQN \
     --stop '{"training_iteration": 2}' \
     --config '{"lr": 1e-3, "schedule_max_timesteps": 100000, "exploration_fraction": 0.1, "exploration_final_eps": 0.02, "dueling": false, "hiddens": [], "model": {"fcnet_hiddens": [64], "fcnet_activation": "relu"}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run DQN \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run APEX \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2, "timesteps_per_iteration": 1000, "gpu": false, "min_iter_time_s": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env FrozenLake-v0 \
     --run DQN \
     --stop '{"training_iteration": 2}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env FrozenLake-v0 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"num_sgd_iter": 10, "sgd_batchsize": 64, "timesteps_per_batch": 1000, "num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env PongDeterministic-v4 \
     --run DQN \
     --stop '{"training_iteration": 2}' \
     --config '{"lr": 1e-4, "schedule_max_timesteps": 2000000, "buffer_size": 10000, "exploration_fraction": 0.1, "exploration_final_eps": 0.01, "sample_batch_size": 4, "learning_starts": 10000, "target_network_update_freq": 1000, "gamma": 0.99, "prioritized_replay": true}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env MontezumaRevenge-v0 \
     --run PPO \
     --stop '{"training_iteration": 2}' \
     --config '{"kl_coeff": 1.0, "num_sgd_iter": 10, "sgd_stepsize": 1e-4, "sgd_batchsize": 64, "timesteps_per_batch": 2000, "num_workers": 1, "model": {"dim": 40, "conv_filters": [[16, [8, 8], 4], [32, [4, 4], 2], [512, [5, 5], 1]]}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env PongDeterministic-v4 \
     --run A3C \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2, "use_pytorch": true, "model": {"use_lstm": false, "grayscale": true, "zero_mean": false, "dim": 80, "channel_major": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run A3C \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2, "use_pytorch": true}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v1 \
     --run A3C \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2, "model": {"use_lstm": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run DQN \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run PG \
     --stop '{"training_iteration": 2}' \
     --config '{"sample_batch_size": 500, "num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run PG \
     --stop '{"training_iteration": 2}' \
     --config '{"sample_batch_size": 500, "num_workers": 1, "model": {"use_lstm": true, "max_seq_len": 100}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run PG \
     --stop '{"training_iteration": 2}' \
     --config '{"sample_batch_size": 500, "num_workers": 1, "num_envs_per_worker": 10}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env Pong-v0 \
     --run PG \
     --stop '{"training_iteration": 2}' \
     --config '{"sample_batch_size": 500, "num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env FrozenLake-v0 \
     --run PG \
     --stop '{"training_iteration": 2}' \
     --config '{"sample_batch_size": 500, "num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env Pendulum-v0 \
     --run DDPG \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run IMPALA \
     --stop '{"training_iteration": 2}' \
     --config '{"gpu": false, "num_workers": 2, "min_iter_time_s": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env CartPole-v0 \
     --run IMPALA \
     --stop '{"training_iteration": 2}' \
     --config '{"gpu": false, "num_workers": 2, "min_iter_time_s": 1, "model": {"use_lstm": true}}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env MountainCarContinuous-v0 \
     --run DDPG \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     rllib train \
     --env MountainCarContinuous-v0 \
     --run DDPG \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/train.py \
     --env Pendulum-v0 \
     --run APEX_DDPG \
@@ -220,69 +220,69 @@ docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     --stop '{"training_iteration": 2}' \
     --config '{"num_workers": 2, "optimizer": {"num_replay_buffer_shards": 1}, "learning_starts": 100, "min_iter_time_s": 1}'
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     sh /ray/test/jenkins_tests/multi_node_tests/test_rllib_eval.sh
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_checkpoint_restore.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_policy_evaluator.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_serving_env.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_lstm.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_multi_agent_env.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/test/test_supported_spaces.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/tune_mnist_ray.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/pbt_example.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/hyperband_example.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/async_hyperband_example.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/tune_mnist_ray_hyperband.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/tune_mnist_async_hyperband.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/hyperopt_example.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/tune/examples/tune_mnist_keras.py \
     --smoke-test
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/examples/legacy_multiagent/multiagent_mountaincar.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/examples/legacy_multiagent/multiagent_pendulum.py
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/examples/multiagent_cartpole.py --num-iters=2
 
-docker run --rm --shm-size=10G --memory=10G $DOCKER_SHA \
+docker run  -e "RAY_USE_XRAY=1" --rm --shm-size=10G --memory=10G $DOCKER_SHA \
     python /ray/python/ray/rllib/examples/multiagent_two_trainers.py --num-iters=2
 
 python3 $ROOT_DIR/multi_node_docker_test.py \