HDFS-14535. The default 8KB buffer in requestFileDescriptors#BufferedOutputStream is causing lots of heap allocation in HBase when using short-circut read

(cherry picked from commit ea3b0a184405c3feca024a560807ea215b6858b9)
(cherry picked from commit 9b61ecfcbeeb1d3c320f775b43c68d3e39a271a3)
(cherry picked from commit b082628e5a39fc1bc9d6094beb9c253cc096c701)

hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/client/impl/BlockReaderFactory.java

@@ -86,6 +86,7 @@ import org.slf4j.LoggerFactory;
 @InterfaceAudience.Private
 public class BlockReaderFactory implements ShortCircuitReplicaCreator {
   static final Logger LOG = LoggerFactory.getLogger(BlockReaderFactory.class);
+  private static final int SMALL_BUFFER_SIZE = 512;
 
   public static class FailureInjector {
     public void injectRequestFileDescriptorsFailure() throws IOException {
@@ -578,7 +579,7 @@ public class BlockReaderFactory implements ShortCircuitReplicaCreator {
           Slot slot) throws IOException {
     ShortCircuitCache cache = clientContext.getShortCircuitCache();
     final DataOutputStream out =
-        new DataOutputStream(new BufferedOutputStream(peer.getOutputStream()));
+        new DataOutputStream(new BufferedOutputStream(peer.getOutputStream(), SMALL_BUFFER_SIZE));
     SlotId slotId = slot == null ? null : slot.getSlotId();
     new Sender(out).requestShortCircuitFds(block, token, slotId, 1,
         failureInjector.getSupportsReceiptVerification());