Merge branch 'master' into unskip_1

.gitignore

@@ -115,10 +115,7 @@ scripts/nodes.txt
 **/MNIST_data/
 **/cifar-10-batches-bin/
 
-# Generated documentation files
-/doc/_build
-/doc/source/_static/thumbs
-/doc/source/tune/generated_guides/
+
 
 # User-specific stuff:
 .idea/**/workspace.xml

benchmarks/benchmark_tests.yaml

@@ -1,8 +1,5 @@
 - name: single_node
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: single_node.yaml
@@ -13,10 +10,7 @@
     script: python single_node/test_single_node.py
 
 - name: object_store
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: object_store.yaml
@@ -27,10 +21,7 @@
     script: python object_store/test_object_store.py
 
 - name: many_actors
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml
@@ -41,10 +32,7 @@
     script: python distributed/test_many_actors.py
 
 - name: many_actors_smoke_test
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed_smoke_test.yaml
@@ -55,10 +43,7 @@
     script: SMOKE_TEST=1 python distributed/test_many_actors.py
 
 - name: many_tasks
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml
@@ -69,10 +54,7 @@
     script: python distributed/test_many_tasks.py --num-tasks=10000
 
 - name: many_tasks_smoke_test
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed_smoke_test.yaml
@@ -83,10 +65,7 @@
     script: python distributed/test_many_tasks.py --num-tasks=100
 
 - name: many_pgs
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml
@@ -97,10 +76,7 @@
     script: python distributed/test_many_pgs.py
 
 - name: many_pgs_smoke_test
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed_smoke_test.yaml
@@ -112,10 +88,7 @@
 
 # NOTE: No smoke test since this shares a script with the many_tasks_smoke_test
 - name: many_nodes
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Alex Wu"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: many_nodes.yaml
@@ -126,10 +99,7 @@
     script: python distributed/test_many_tasks.py --num-tasks=1000
 
 - name: many_tasks_redis_ha
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Yi Cheng"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml
@@ -146,10 +116,7 @@
   stable: false
 
 - name: many_actors_redis_ha
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Yi Cheng"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml
@@ -166,10 +133,7 @@
   stable: false
 
 - name: many_nodes_redis_ha
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Yi Cheng"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: many_nodes.yaml
@@ -186,10 +150,7 @@
   stable: false
 
 - name: many_pgs_redis_ha
-  owner:
-    mail: "core@anyscale.com"
-    slack: "@Yi Cheng"
-
+  team: core
   cluster:
     app_config: app_config.yaml
     compute_template: distributed.yaml

dashboard/modules/node/node_head.py

@@ -261,9 +261,11 @@ def process_log_batch(log_batch):
 
         if self._dashboard_head.gcs_log_subscriber:
             while True:
-                log_batch = await \
-                    self._dashboard_head.gcs_log_subscriber.poll()
                 try:
+                    log_batch = await \
+                        self._dashboard_head.gcs_log_subscriber.poll()
+                    if log_batch is None:
+                        continue
                     process_log_batch(log_batch)
                 except Exception:
                     logger.exception("Error receiving log from GCS.")
@@ -304,9 +306,11 @@ def process_error(error_data):
 
         if self._dashboard_head.gcs_error_subscriber:
             while True:
-                _, error_data = await \
-                    self._dashboard_head.gcs_error_subscriber.poll()
                 try:
+                    _, error_data = await \
+                        self._dashboard_head.gcs_error_subscriber.poll()
+                    if error_data is None:
+                        continue
                     process_error(error_data)
                 except Exception:
                     logger.exception("Error receiving error info from GCS.")

dashboard/modules/reporter/reporter_agent.py

@@ -29,6 +29,8 @@
 
 logger = logging.getLogger(__name__)
 
+enable_gpu_usage_check = True
+
 # Are we in a K8s pod?
 IN_KUBERNETES_POD = "KUBERNETES_SERVICE_HOST" in os.environ
 
@@ -202,14 +204,26 @@ def _get_cpu_percent():
 
     @staticmethod
     def _get_gpu_usage():
-        if gpustat is None:
+        global enable_gpu_usage_check
+        if gpustat is None or not enable_gpu_usage_check:
             return []
         gpu_utilizations = []
         gpus = []
         try:
             gpus = gpustat.new_query().gpus
         except Exception as e:
             logger.debug(f"gpustat failed to retrieve GPU information: {e}")
+
+            # gpustat calls pynvml.nvmlInit()
+            # On machines without GPUs, this can run subprocesses that spew to
+            # stderr. Then with log_to_driver=True, we get log spew from every
+            # single raylet. To avoid this, disable the GPU usage check on
+            # certain errors.
+            # https://github.com/ray-project/ray/issues/14305
+            # https://github.com/ray-project/ray/pull/21686
+            if type(e).__name__ == "NVMLError_DriverNotLoaded":
+                enable_gpu_usage_check = False
+
         for gpu in gpus:
             # Note the keys in this dict have periods which throws
             # off javascript so we change .s to _s

doc/.gitignore

@@ -1 +1,8 @@
-auto_examples/
+# Generated documentation files
+_build
+source/_static/thumbs
+
+source/ray-core/examples/
+source/ray-tune/tutorials/
+source/ray-tune/generated_guides/
+source/ray-data/examples/

doc/BUILD

@@ -8,8 +8,8 @@
 py_test(
     name = "dask_xgboost",
     size = "medium",
-    main = "examples/dask_xgboost/dask_xgboost.py",
-    srcs = ["examples/dask_xgboost/dask_xgboost.py"],
+    main = "source/ray-core/_examples/dask_xgboost/dask_xgboost.py",
+    srcs = ["source/ray-core/_examples/dask_xgboost/dask_xgboost.py"],
     tags = ["exclusive", "team:ml", "py37"],
     args = ["--smoke-test", "--address ''", "--num-actors 4",
             "--cpus-per-actor 1", "--num-actors-inference 4",
@@ -20,8 +20,8 @@ py_test(
 py_test(
     name = "modin_xgboost",
     size = "medium",
-    main = "examples/modin_xgboost/modin_xgboost.py",
-    srcs = ["examples/modin_xgboost/modin_xgboost.py"],
+    main = "source/ray-core/_examples/modin_xgboost/modin_xgboost.py",
+    srcs = ["source/ray-core/_examples/modin_xgboost/modin_xgboost.py"],
     tags = ["exclusive", "team:ml", "py37"],
     args = ["--smoke-test", "--address ''", "--num-actors 4",
             "--cpus-per-actor 1", "--num-actors-inference 4",
@@ -31,77 +31,77 @@ py_test(
 py_test(
     name = "big_data_ingestion",
     size = "small",
-    srcs = ["source/data/_examples/big_data_ingestion.py"],
+    srcs = ["source/ray-data/_examples/big_data_ingestion.py"],
     tags = ["exclusive", "team:core", "py37"]
 )
 
 py_test(
     name = "datasets_train",
     size = "medium",
-    srcs = ["examples/datasets_train/datasets_train.py"],
+    srcs = ["source/ray-core/_examples/datasets_train/datasets_train.py"],
     tags = ["exclusive", "team:ml", "py37", "datasets_train"],
     args = ["--smoke-test", "--num-workers=2", "--use-gpu"]
 )
 
 py_test(
     name = "plot_hyperparameter",
     size = "small",
-    srcs = ["examples/plot_hyperparameter.py"],
+    srcs = ["source/ray-core/_examples/plot_hyperparameter.py"],
     tags = ["exclusive", "team:ml"]
 )
 
 py_test(
     name = "plot_parameter_server",
     size = "medium",
-    srcs = ["examples/plot_parameter_server.py"],
+    srcs = ["source/ray-core/_examples/plot_parameter_server.py"],
     tags = ["exclusive", "team:ml"]
 )
 
 py_test(
     name = "plot_pong_example",
     size = "large",
-    srcs = ["examples/plot_pong_example.py"],
+    srcs = ["source/ray-core/_examples/plot_pong_example.py"],
     tags = ["exclusive", "team:ml"]
 )
 
 py_test(
     name = "progress_bar",
     size = "small",
-    srcs = ["examples/progress_bar.py"],
+    srcs = ["source/ray-core/_examples/progress_bar.py"],
     tags = ["exclusive", "team:ml"]
 )
 
 
-# Directory: examples/doc_code
+# Directory: source/ray-core/_examples/doc_code
 py_test(
     name = "doc_code_tf_example",
     size = "small",
-    main = "examples/doc_code/tf_example.py",
-    srcs = ["examples/doc_code/tf_example.py"],
+    main = "source/ray-core/_examples/doc_code/tf_example.py",
+    srcs = ["source/ray-core/_examples/doc_code/tf_example.py"],
     tags = ["exclusive", "tf", "team:ml"]
 )
 
 py_test(
     name = "doc_code_torch_example",
     size = "small",
-    main = "examples/doc_code/torch_example.py",
-    srcs = ["examples/doc_code/torch_example.py"],
+    main = "source/ray-core/_examples/doc_code/torch_example.py",
+    srcs = ["source/ray-core/_examples/doc_code/torch_example.py"],
     tags = ["exclusive", "pytorch", "team:ml"]
 )
 
 py_test(
     name = "doc_code_metrics_example",
     size = "small",
-    main = "examples/doc_code/metrics_example.py",
-    srcs = ["examples/doc_code/metrics_example.py"],
+    main = "source/ray-core/_examples/doc_code/metrics_example.py",
+    srcs = ["source/ray-core/_examples/doc_code/metrics_example.py"],
     tags = ["exclusive", "team:serve"]
 )
 
 py_test(
     name = "doc_code_runtime_env_example",
     size = "small",
-    main = "examples/doc_code/runtime_env_example.py",
-    srcs = ["examples/doc_code/runtime_env_example.py"],
+    main = "source/ray-core/_examples/doc_code/runtime_env_example.py",
+    srcs = ["source/ray-core/_examples/doc_code/runtime_env_example.py"],
     tags = ["exclusive", "post_wheel_build", "team:serve"]
 )
 
@@ -113,17 +113,17 @@ py_test(
 py_test(
     name = "tune_sklearn",
     size = "medium",
-    main = "source/tune/_tutorials/tune-sklearn.py",
-    srcs = ["source/tune/_tutorials/tune-sklearn.py"],
+    main = "source/ray-tune/_tutorials/tune-sklearn.py",
+    srcs = ["source/ray-tune/_tutorials/tune-sklearn.py"],
     tags = ["exclusive", "example", "team:ml"],
     args = ["--smoke-test"]
 )
 
 py_test(
     name = "tune_serve_integration_mnist",
     size = "medium",
-    main = "source/tune/_tutorials/tune-serve-integration-mnist.py",
-    srcs = ["source/tune/_tutorials/tune-serve-integration-mnist.py"],
+    main = "source/ray-tune/_tutorials/tune-serve-integration-mnist.py",
+    srcs = ["source/ray-tune/_tutorials/tune-serve-integration-mnist.py"],
     tags = ["exclusive", "example", "team:ml"],
     args = ["--smoke-test", "--from_scratch", "--day 0"]
 )