HDFS-11047. Remove deep copies of FinalizedReplica to alleviate heap consumption on DataNode. Contributed by Xiaobing Zhou

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DirectoryScanner.java

@@ -22,6 +22,7 @@ import java.io.File;
 import java.io.FilenameFilter;
 import java.io.IOException;
 import java.util.Arrays;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.LinkedList;
 import java.util.List;
@@ -398,14 +399,13 @@ public class DirectoryScanner implements Runnable {
         diffs.put(bpid, diffRecord);
         
         statsRecord.totalBlocks = blockpoolReport.length;
-        List<ReplicaInfo> bl = dataset.getFinalizedBlocks(bpid);
-        ReplicaInfo[] memReport = bl.toArray(new ReplicaInfo[bl.size()]);
-        Arrays.sort(memReport); // Sort based on blockId
+        final List<ReplicaInfo> bl = dataset.getFinalizedBlocks(bpid);
+        Collections.sort(bl); // Sort based on blockId
   
         int d = 0; // index for blockpoolReport
         int m = 0; // index for memReprot
-        while (m < memReport.length && d < blockpoolReport.length) {
-          ReplicaInfo memBlock = memReport[m];
+        while (m < bl.size() && d < blockpoolReport.length) {
+          ReplicaInfo memBlock = bl.get(m);
           ScanInfo info = blockpoolReport[d];
           if (info.getBlockId() < memBlock.getBlockId()) {
             if (!dataset.isDeletingBlock(bpid, info.getBlockId())) {
@@ -452,8 +452,8 @@ public class DirectoryScanner implements Runnable {
             ++m;
           }
         }
-        while (m < memReport.length) {
-          ReplicaInfo current = memReport[m++];
+        while (m < bl.size()) {
+          ReplicaInfo current = bl.get(m++);
           addDifference(diffRecord, statsRecord,
                         current.getBlockId(), current.getVolume());
         }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/fsdataset/FsDatasetSpi.java

@@ -229,7 +229,16 @@ public interface FsDatasetSpi<V extends FsVolumeSpi> extends FSDatasetMBean {
    */
   VolumeFailureSummary getVolumeFailureSummary();
 
-  /** @return a list of finalized blocks for the given block pool. */
+  /**
+   * Gets a list of references to the finalized blocks for the given block pool.
+   * <p>
+   * Callers of this function should call
+   * {@link FsDatasetSpi#acquireDatasetLock} to avoid blocks' status being
+   * changed during list iteration.
+   * </p>
+   * @return a list of references to the finalized blocks for the given block
+   *         pool.
+   */
   List<ReplicaInfo> getFinalizedBlocks(String bpid);
 
   /** @return a list of finalized blocks for the given block pool. */

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/fsdataset/impl/FsDatasetImpl.java

@@ -1714,17 +1714,23 @@ class FsDatasetImpl implements FsDatasetSpi<FsVolumeImpl> {
   }
 
   /**
-   * Get the list of finalized blocks from in-memory blockmap for a block pool.
+   * Gets a list of references to the finalized blocks for the given block pool.
+   * <p>
+   * Callers of this function should call
+   * {@link FsDatasetSpi#acquireDatasetLock} to avoid blocks' status being
+   * changed during list iteration.
+   * </p>
+   * @return a list of references to the finalized blocks for the given block
+   *         pool.
    */
   @Override
   public List<ReplicaInfo> getFinalizedBlocks(String bpid) {
     try (AutoCloseableLock lock = datasetLock.acquire()) {
-      ArrayList<ReplicaInfo> finalized =
-          new ArrayList<ReplicaInfo>(volumeMap.size(bpid));
+      final List<ReplicaInfo> finalized = new ArrayList<ReplicaInfo>(
+          volumeMap.size(bpid));
       for (ReplicaInfo b : volumeMap.replicas(bpid)) {
         if (b.getState() == ReplicaState.FINALIZED) {
-          finalized.add(new ReplicaBuilder(ReplicaState.FINALIZED)
-              .from(b).build());
+          finalized.add(b);
         }
       }
       return finalized;