You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2020/09/01 15:07:34 UTC
[GitHub] [spark] tgravescs commented on a change in pull request #28618: [SPARK-31801][API][SHUFFLE] Register map output metadata
tgravescs commented on a change in pull request #28618:
URL: https://github.com/apache/spark/pull/28618#discussion_r481159028
##########
File path: core/src/main/java/org/apache/spark/shuffle/api/metadata/ShuffleOutputTracker.java
##########
@@ -0,0 +1,56 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle.api.metadata;
+
+/**
+ * :: Private ::
+ *
+ * A plugin that can monitor the storage of shuffle data from map tasks, and can provide
+ * metadata to shuffle readers to aid their reading of shuffle blocks in reduce tasks.
+ * <p>
+ * {@link MapOutputMetadata} instances provided from the plugin tree's implementation of
+ * {@link org.apache.spark.shuffle.api.ShuffleMapOutputWriter} are sent to
Review comment:
sent to ... .what?
##########
File path: core/src/main/java/org/apache/spark/shuffle/api/metadata/NoOpShuffleOutputTracker.java
##########
@@ -0,0 +1,34 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle.api.metadata;
+
+public class NoOpShuffleOutputTracker implements ShuffleOutputTracker {
Review comment:
add description
##########
File path: core/src/main/java/org/apache/spark/shuffle/api/metadata/ShuffleOutputTracker.java
##########
@@ -0,0 +1,56 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle.api.metadata;
+
+/**
+ * :: Private ::
+ *
+ * A plugin that can monitor the storage of shuffle data from map tasks, and can provide
Review comment:
assume this is on the driver, might be nice just to mention.
##########
File path: core/src/main/java/org/apache/spark/shuffle/api/metadata/ShuffleOutputTracker.java
##########
@@ -0,0 +1,56 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle.api.metadata;
+
+/**
+ * :: Private ::
Review comment:
why are we calling all of this private if we want people to implement? Wouldn't developer api and evolving be better?
##########
File path: core/src/main/java/org/apache/spark/shuffle/api/metadata/ShuffleOutputTracker.java
##########
@@ -0,0 +1,56 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle.api.metadata;
+
+/**
+ * :: Private ::
+ *
+ * A plugin that can monitor the storage of shuffle data from map tasks, and can provide
+ * metadata to shuffle readers to aid their reading of shuffle blocks in reduce tasks.
+ * <p>
+ * {@link MapOutputMetadata} instances provided from the plugin tree's implementation of
+ * {@link org.apache.spark.shuffle.api.ShuffleMapOutputWriter} are sent to
+ * <p>
+ * Implementations MUST be thread-safe. Spark will invoke methods in this module in parallel.
+ */
+public interface ShuffleOutputTracker {
+
+ /**
+ * Called when a new shuffle stage is going to be run.
+ */
+ void registerShuffle(int shuffleId);
Review comment:
it would be nice to have full java doc comments explaining parameter on all of these. Most are obvious but its still better especially if people are meant to implement.
##########
File path: core/src/main/scala/org/apache/spark/shuffle/MemoizingShuffleDataIO.scala
##########
@@ -0,0 +1,45 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle
+
+import scala.collection.JavaConverters._
+
+import org.apache.spark.SparkEnv
+import org.apache.spark.shuffle.api.{ShuffleDataIO, ShuffleDriverComponents, ShuffleExecutorComponents}
+
+/**
+ * Thin wrapper around {@link ShuffleDataIO} that ensures the given components are
+ * only initialized once and providing the same instance each time.
+ * <p>
+ * Used to ensure the SparkEnv only instantiates the given components once lazily
+ * and then reuses them throughout the lifetime of the SparkEnv.
+ */
+class MemoizingShuffleDataIO(delegate: ShuffleDataIO) {
Review comment:
I don't completely follow why we need this, especially on the driver side, can you explain? Are we creating the SparkEnv multiple times or are you worried about the user doing it, or just in case? Is it more for the lazy side of things. Just want to understand.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org