HDFS-9412. getBlocks occupies FSLock and takes too long to complete. Contributed by He Tianyi.

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java

@@ -305,6 +305,14 @@ public class BlockManager implements BlockStatsMXBean {
    * processed again after aquiring lock again.
    */
   private int numBlocksPerIteration;
+
+  /**
+   * Minimum size that a block can be sent to Balancer through getBlocks.
+   * And after HDFS-8824, the small blocks are unused anyway, so there's no
+   * point to send them to balancer.
+   */
+  private long getBlocksMinBlockSize = -1;
+
   /**
    * Progress of the Reconstruction queues initialisation.
    */
@@ -414,6 +422,9 @@ public class BlockManager implements BlockStatsMXBean {
     this.numBlocksPerIteration = conf.getInt(
         DFSConfigKeys.DFS_BLOCK_MISREPLICATION_PROCESSING_LIMIT,
         DFSConfigKeys.DFS_BLOCK_MISREPLICATION_PROCESSING_LIMIT_DEFAULT);
+    this.getBlocksMinBlockSize = conf.getLongBytes(
+        DFSConfigKeys.DFS_BALANCER_GETBLOCKS_MIN_BLOCK_SIZE_KEY,
+        DFSConfigKeys.DFS_BALANCER_GETBLOCKS_MIN_BLOCK_SIZE_DEFAULT);
     this.blockReportLeaseManager = new BlockReportLeaseManager(conf);
 
     bmSafeMode = new BlockManagerSafeMode(this, namesystem, haEnabled, conf);
@@ -1179,6 +1190,9 @@ public class BlockManager implements BlockStatsMXBean {
     while(totalSize<size && iter.hasNext()) {
       curBlock = iter.next();
       if(!curBlock.isComplete())  continue;
+      if (curBlock.getNumBytes() < getBlocksMinBlockSize) {
+        continue;
+      }
       totalSize += addBlock(curBlock, results);
     }
     if(totalSize<size) {
@@ -1186,6 +1200,9 @@ public class BlockManager implements BlockStatsMXBean {
       for(int i=0; i<startBlock&&totalSize<size; i++) {
         curBlock = iter.next();
         if(!curBlock.isComplete())  continue;
+        if (curBlock.getNumBytes() < getBlocksMinBlockSize) {
+          continue;
+        }
         totalSize += addBlock(curBlock, results);
       }
     }

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestGetBlocks.java

@@ -179,11 +179,15 @@ public class TestGetBlocks {
     final int DEFAULT_BLOCK_SIZE = 1024;
 
     CONF.setLong(DFSConfigKeys.DFS_BLOCK_SIZE_KEY, DEFAULT_BLOCK_SIZE);
+    CONF.setLong(DFSConfigKeys.DFS_BALANCER_GETBLOCKS_MIN_BLOCK_SIZE_KEY,
+      DEFAULT_BLOCK_SIZE);
+
     MiniDFSCluster cluster = new MiniDFSCluster.Builder(CONF).numDataNodes(
         REPLICATION_FACTOR).build();
     try {
       cluster.waitActive();
-      long fileLen = 2 * DEFAULT_BLOCK_SIZE;
+      // the third block will not be visible to getBlocks
+      long fileLen = 2 * DEFAULT_BLOCK_SIZE + 1;
       DFSTestUtil.createFile(cluster.getFileSystem(), new Path("/tmp.txt"),
           fileLen, REPLICATION_FACTOR, 0L);
 
@@ -196,7 +200,7 @@ public class TestGetBlocks {
             DFSUtilClient.getNNAddress(CONF), CONF);
         locatedBlocks = dfsclient.getNamenode()
             .getBlockLocations("/tmp.txt", 0, fileLen).getLocatedBlocks();
-        assertEquals(2, locatedBlocks.size());
+        assertEquals(3, locatedBlocks.size());
         notWritten = false;
         for (int i = 0; i < 2; i++) {
           dataNodes = locatedBlocks.get(i).getLocations();