HADOOP-10591. Compression codecs must used pooled direct buffers or deallocate direct buffers when stream is closed (cmccabe)

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1611427 13f79535-47bb-0310-9956-ffa450edef68

hadoop-common-project/hadoop-common/CHANGES.txt

@@ -51,6 +51,9 @@ Release 2.6.0 - UNRELEASED
     HADOOP-9921. daemon scripts should remove pid file on stop call after stop
     or process is found not running ( vinayakumarb )
 
+    HADOOP-10591.  Compression codecs must used pooled direct buffers or
+    deallocate direct buffers when stream is closed (cmccabe)
+
 Release 2.5.0 - UNRELEASED
 
   INCOMPATIBLE CHANGES

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/BZip2Codec.java

@@ -100,7 +100,8 @@ public class BZip2Codec implements Configurable, SplittableCompressionCodec {
   @Override
   public CompressionOutputStream createOutputStream(OutputStream out)
       throws IOException {
-    return createOutputStream(out, createCompressor());
+    return CompressionCodec.Util.
+        createOutputStreamWithCodecPool(this, conf, out);
   }
 
   /**
@@ -153,7 +154,8 @@ public class BZip2Codec implements Configurable, SplittableCompressionCodec {
   @Override
   public CompressionInputStream createInputStream(InputStream in)
       throws IOException {
-    return createInputStream(in, createDecompressor());
+    return CompressionCodec.Util.
+        createInputStreamWithCodecPool(this, conf, in);
   }
 
   /**

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/CompressionCodec.java

@@ -24,6 +24,7 @@ import java.io.OutputStream;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
+import org.apache.hadoop.conf.Configuration;
 
 /**
  * This class encapsulates a streaming compression/decompression pair.
@@ -113,4 +114,58 @@ public interface CompressionCodec {
    * @return the extension including the '.'
    */
   String getDefaultExtension();
+
+  static class Util {
+    /**
+     * Create an output stream with a codec taken from the global CodecPool.
+     *
+     * @param codec       The codec to use to create the output stream.
+     * @param conf        The configuration to use if we need to create a new codec.
+     * @param out         The output stream to wrap.
+     * @return            The new output stream
+     * @throws IOException
+     */
+    static CompressionOutputStream createOutputStreamWithCodecPool(
+        CompressionCodec codec, Configuration conf, OutputStream out)
+        throws IOException {
+      Compressor compressor = CodecPool.getCompressor(codec, conf);
+      CompressionOutputStream stream = null;
+      try {
+        stream = codec.createOutputStream(out, compressor);
+      } finally {
+        if (stream == null) {
+          CodecPool.returnCompressor(compressor);
+        } else {
+          stream.setTrackedCompressor(compressor);
+        }
+      }
+      return stream;
+    }
+
+    /**
+     * Create an input stream with a codec taken from the global CodecPool.
+     *
+     * @param codec       The codec to use to create the input stream.
+     * @param conf        The configuration to use if we need to create a new codec.
+     * @param in          The input stream to wrap.
+     * @return            The new input stream
+     * @throws IOException
+     */
+    static CompressionInputStream createInputStreamWithCodecPool(
+        CompressionCodec codec,  Configuration conf, InputStream in)
+          throws IOException {
+      Decompressor decompressor = CodecPool.getDecompressor(codec);
+      CompressionInputStream stream = null;
+      try {
+        stream = codec.createInputStream(in, decompressor);
+      } finally {
+        if (stream == null) {
+          CodecPool.returnDecompressor(decompressor);
+        } else {
+          stream.setTrackedDecompressor(decompressor);
+        }
+      }
+      return stream;
+    }
+  }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/CompressionInputStream.java

@@ -41,6 +41,8 @@ public abstract class CompressionInputStream extends InputStream implements Seek
   protected final InputStream in;
   protected long maxAvailableData = 0L;
 
+  private Decompressor trackedDecompressor;
+
   /**
    * Create a compression input stream that reads
    * the decompressed bytes from the given stream.
@@ -58,6 +60,10 @@ public abstract class CompressionInputStream extends InputStream implements Seek
   @Override
   public void close() throws IOException {
     in.close();
+    if (trackedDecompressor != null) {
+      CodecPool.returnDecompressor(trackedDecompressor);
+      trackedDecompressor = null;
+    }
   }
   
   /**
@@ -112,4 +118,8 @@ public abstract class CompressionInputStream extends InputStream implements Seek
   public boolean seekToNewSource(long targetPos) throws UnsupportedOperationException {
     throw new UnsupportedOperationException();
   }
+
+  void setTrackedDecompressor(Decompressor decompressor) {
+    trackedDecompressor = decompressor;
+  }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/CompressionOutputStream.java

@@ -35,6 +35,12 @@ public abstract class CompressionOutputStream extends OutputStream {
    */
   protected final OutputStream out;
 
+  /**
+   * If non-null, this is the Compressor object that we should call
+   * CodecPool#returnCompressor on when this stream is closed.
+   */
+  private Compressor trackedCompressor;
+
   /**
    * Create a compression output stream that writes
    * the compressed bytes to the given stream.
@@ -44,10 +50,18 @@ public abstract class CompressionOutputStream extends OutputStream {
     this.out = out;
   }
 
+  void setTrackedCompressor(Compressor compressor) {
+    trackedCompressor = compressor;
+  }
+
   @Override
   public void close() throws IOException {
     finish();
     out.close();
+    if (trackedCompressor != null) {
+      CodecPool.returnCompressor(trackedCompressor);
+      trackedCompressor = null;
+    }
   }
   
   @Override

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/DefaultCodec.java

@@ -51,14 +51,8 @@ public class DefaultCodec implements Configurable, CompressionCodec, DirectDecom
   @Override
   public CompressionOutputStream createOutputStream(OutputStream out) 
   throws IOException {
-    // This may leak memory if called in a loop. The createCompressor() call
+    return CompressionCodec.Util.
-    // may cause allocation of an untracked direct-backed buffer if native
+        createOutputStreamWithCodecPool(this, conf, out);
-    // libs are being used (even if you close the stream).  A Compressor
-    // object should be reused between successive calls.
-    LOG.warn("DefaultCodec.createOutputStream() may leak memory. "
-        + "Create a compressor first.");
-    return new CompressorStream(out, createCompressor(), 
-                                conf.getInt("io.file.buffer.size", 4*1024));
   }
 
   @Override
@@ -82,8 +76,8 @@ public class DefaultCodec implements Configurable, CompressionCodec, DirectDecom
   @Override
   public CompressionInputStream createInputStream(InputStream in) 
   throws IOException {
-    return new DecompressorStream(in, createDecompressor(),
+    return CompressionCodec.Util.
-                                  conf.getInt("io.file.buffer.size", 4*1024));
+        createInputStreamWithCodecPool(this, conf, in);
   }
 
   @Override

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/GzipCodec.java

@@ -104,10 +104,11 @@ public class GzipCodec extends DefaultCodec {
   @Override
   public CompressionOutputStream createOutputStream(OutputStream out) 
     throws IOException {
-    return (ZlibFactory.isNativeZlibLoaded(conf)) ?
+    if (!ZlibFactory.isNativeZlibLoaded(conf)) {
-               new CompressorStream(out, createCompressor(),
+      return new GzipOutputStream(out);
-                                    conf.getInt("io.file.buffer.size", 4*1024)) :
+    }
-               new GzipOutputStream(out);
+    return CompressionCodec.Util.
+        createOutputStreamWithCodecPool(this, conf, out);
   }
   
   @Override
@@ -138,7 +139,8 @@ public class GzipCodec extends DefaultCodec {
   @Override
   public CompressionInputStream createInputStream(InputStream in)
       throws IOException {
-    return createInputStream(in, null);
+    return CompressionCodec.Util.
+        createInputStreamWithCodecPool(this, conf, in);
   }
 
   @Override

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/Lz4Codec.java

@@ -84,7 +84,8 @@ public class Lz4Codec implements Configurable, CompressionCodec {
   @Override
   public CompressionOutputStream createOutputStream(OutputStream out)
       throws IOException {
-    return createOutputStream(out, createCompressor());
+    return CompressionCodec.Util.
+        createOutputStreamWithCodecPool(this, conf, out);
   }
 
   /**
@@ -157,7 +158,8 @@ public class Lz4Codec implements Configurable, CompressionCodec {
   @Override
   public CompressionInputStream createInputStream(InputStream in)
       throws IOException {
-    return createInputStream(in, createDecompressor());
+    return CompressionCodec.Util.
+        createInputStreamWithCodecPool(this, conf, in);
   }
 
   /**

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/compress/SnappyCodec.java

@@ -95,7 +95,8 @@ public class SnappyCodec implements Configurable, CompressionCodec, DirectDecomp
   @Override
   public CompressionOutputStream createOutputStream(OutputStream out)
       throws IOException {
-    return createOutputStream(out, createCompressor());
+    return CompressionCodec.Util.
+        createOutputStreamWithCodecPool(this, conf, out);
   }
 
   /**
@@ -158,7 +159,8 @@ public class SnappyCodec implements Configurable, CompressionCodec, DirectDecomp
   @Override
   public CompressionInputStream createInputStream(InputStream in)
       throws IOException {
-    return createInputStream(in, createDecompressor());
+    return CompressionCodec.Util.
+        createInputStreamWithCodecPool(this, conf, in);
   }
 
   /**