HDFS-12409. Add metrics of execution time of different stages in EC recovery task. (Lei (Eddy) Xu)

(cherry picked from commit 73aed34dff)

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

@@ -22,6 +22,7 @@ import java.nio.ByteBuffer;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.hdfs.server.datanode.metrics.DataNodeMetrics;
+import org.apache.hadoop.util.Time;
 
 /**
  * StripedBlockReconstructor reconstruct one or more missed striped block in
@@ -83,18 +84,28 @@ class StripedBlockReconstructor extends StripedReconstructor
       final int toReconstructLen =
           (int) Math.min(getStripedReader().getBufferSize(), remaining);
 
+      long start = Time.monotonicNow();
       // step1: read from minimum source DNs required for reconstruction.
       // The returned success list is the source DNs we do real read from
       getStripedReader().readMinimumSources(toReconstructLen);
+      long readEnd = Time.monotonicNow();
 
       // step2: decode to reconstruct targets
       reconstructTargets(toReconstructLen);
+      long decodeEnd = Time.monotonicNow();
 
       // step3: transfer data
       if (stripedWriter.transferData2Targets() == 0) {
         String error = "Transfer failed for all targets.";
         throw new IOException(error);
       }
+      long writeEnd = Time.monotonicNow();
+
+      // Only the succeed reconstructions are recorded.
+      final DataNodeMetrics metrics = getDatanode().getMetrics();
+      metrics.incrECReconstructionReadTime(readEnd - start);
+      metrics.incrECReconstructionDecodingTime(decodeEnd - readEnd);
+      metrics.incrECReconstructionWriteTime(writeEnd - decodeEnd);
 
       updatePositionInBlock(toReconstructLen);
 

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

@@ -151,6 +151,12 @@ public class DataNodeMetrics {
   MutableCounterLong ecReconstructionBytesWritten;
   @Metric("Bytes remote read by erasure coding worker")
   MutableCounterLong ecReconstructionRemoteBytesRead;
+  @Metric("Milliseconds spent on read by erasure coding worker")
+  private MutableCounterLong ecReconstructionReadTimeMillis;
+  @Metric("Milliseconds spent on decoding by erasure coding worker")
+  private MutableCounterLong ecReconstructionDecodingTimeMillis;
+  @Metric("Milliseconds spent on write by erasure coding worker")
+  private MutableCounterLong ecReconstructionWriteTimeMillis;
 
   final MetricsRegistry registry = new MetricsRegistry("datanode");
   final String name;
@@ -503,4 +509,16 @@ public class DataNodeMetrics {
   public void incrECReconstructionBytesWritten(long bytes) {
     ecReconstructionBytesWritten.incr(bytes);
   }
+
+  public void incrECReconstructionReadTime(long millis) {
+    ecReconstructionReadTimeMillis.incr(millis);
+  }
+
+  public void incrECReconstructionWriteTime(long millis) {
+    ecReconstructionWriteTimeMillis.incr(millis);
+  }
+
+  public void incrECReconstructionDecodingTime(long millis) {
+    ecReconstructionDecodingTimeMillis.incr(millis);
+  }
 }

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/datanode/TestDataNodeErasureCodingMetrics.java

@@ -90,6 +90,10 @@ public class TestDataNodeErasureCodingMetrics {
 
   @Test(timeout = 120000)
   public void testFullBlock() throws Exception {
+    Assert.assertEquals(0, getLongMetric("EcReconstructionReadTimeMillis"));
+    Assert.assertEquals(0, getLongMetric("EcReconstructionDecodingTimeMillis"));
+    Assert.assertEquals(0, getLongMetric("EcReconstructionWriteTimeMillis"));
+
     doTest("/testEcMetrics", blockGroupSize, 0);
 
     Assert.assertEquals("EcReconstructionTasks should be ",
@@ -103,6 +107,9 @@ public class TestDataNodeErasureCodingMetrics {
         blockSize, getLongMetric("EcReconstructionBytesWritten"));
     Assert.assertEquals("EcReconstructionRemoteBytesRead should be ",
         0, getLongMetricWithoutCheck("EcReconstructionRemoteBytesRead"));
+    Assert.assertTrue(getLongMetric("EcReconstructionReadTimeMillis") > 0);
+    Assert.assertTrue(getLongMetric("EcReconstructionDecodingTimeMillis") > 0);
+    Assert.assertTrue(getLongMetric("EcReconstructionWriteTimeMillis") > 0);
   }
 
   // A partial block, reconstruct the partial block