vanzin commented on a change in pull request #25007: 
[SPARK-28209][CORE][SHUFFLE] Proposed new shuffle writer API 
URL: https://github.com/apache/spark/pull/25007#discussion_r307044685
 
 

 ##########
 File path: 
core/src/main/java/org/apache/spark/shuffle/api/TransferrableWritableByteChannel.java
 ##########
 @@ -0,0 +1,59 @@
+/*
+ * 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;
+
+import java.io.Closeable;
+import java.io.IOException;
+
+import java.nio.channels.FileChannel;
+import java.nio.channels.WritableByteChannel;
+import org.apache.spark.annotation.Private;
+
+/**
+ * :: Private ::
+ * Represents an output byte channel that can copy bytes from input file 
channels to some
+ * arbitrary storage system.
+ * <p>
+ * This API is provided for advanced users who can transfer bytes from a file 
channel to
+ * some output sink without copying data into memory. Most users should not 
need to use
+ * this functionality; this is primarily provided for the built-in shuffle 
storage backends
+ * that persist shuffle files on local disk.
+ * <p>
+ * For a simpler alternative, see {@link ShufflePartitionWriter}.
+ *
+ * @since 3.0.0
+ */
+@Private
+public interface TransferrableWritableByteChannel extends Closeable {
+
+  /**
+   * Copy all bytes from the source readable byte channel into this byte 
channel.
+   * <p>
+   * This method should block until all of the bytes from the source (that is, 
up until
+   * numBytesToTransfer) are available in the output storage layer.
+   *
+   * @param source File to transfer bytes from. Do not call anything on this 
channel other than
 
 Review comment:
   > Basically this comes from a need to block the close method from closing 
the underlying channel
   
   I see. Still this is a really convoluted interface, which makes me more 
strongly prefer that this be hidden as an optimization in the internal, local 
file implementation, instead of being exposed in the public API.
   
   If really exposing this in the public API, I think it would be better to 
have an interface that makes it clear what's the intent here: provide a 
`FileChannel` (and not any other kind of channel) that, depending on the 
plugin, should not be closed. So something like:
   
   ```
   interface FileChannelWrapper {
   
     FileChannel openChannel();
   
     boolean shouldClose();
   
   }
   ```
   
   Plugins by default would just return a `null` wrapper, and then the code 
calling this could do:
   
   ```
   FileChannel channel = transferToEnabled ? writer.fileChannelWrapper() : null;
   if (channel != null) {
     Utils.copyBlahBlahBlah();
     if (channel.shouldClose()) channel.close();
   } else {
     // non-nio code path.
   }
   ```
   
   It's not great, but it's (i) less code than you have here and (ii) much 
clearer what's the actual intent of this API, without the restrictions about 
what methods should and should not be called.

----------------------------------------------------------------
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:
us...@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org

Reply via email to