Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merging trunk to HDFS-2802 branch. 

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/HDFS-2802@1416603 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Suresh Srinivas 2012-12-03 18:04:51 +00:00
parent 6ad4752853 e5d54ac89a
commit d500d59cbe
42 changed files with 1469 additions and 333 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
hadoop-common-project/hadoop-common/CHANGES.txt
View File

@ -306,6 +306,10 @@ Release 2.0.3-alpha - Unreleased
HADOOP-9020. Add a SASL PLAIN server (daryn via bobby)
HADOOP-9090. Support on-demand publish of metrics. (Mostafa Elhemali via
suresh)
IMPROVEMENTS
HADOOP-8789. Tests setLevel(Level.OFF) should be Level.ERROR.

11
hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/metrics2/MetricsSystem.java
View File

@ -90,6 +90,17 @@ public abstract class MetricsSystem implements MetricsSystemMXBean {
*/
public abstract void register(Callback callback);
/**
* Requests an immediate publish of all metrics from sources to sinks.
*
* This is a "soft" request: the expectation is that a best effort will be
* done to synchronously snapshot the metrics from all the sources and put
* them in all the sinks (including flushing the sinks) before returning to
* the caller. If this can't be accomplished in reasonable time it's OK to
* return to the caller before everything is done.
*/
public abstract void publishMetricsNow();
/**
* Shutdown the metrics system completely (usually during server shutdown.)
* The MetricsSystemMXBean will be unregistered.

46
hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/metrics2/impl/MetricsSinkAdapter.java
View File

@ -19,6 +19,7 @@
package org.apache.hadoop.metrics2.impl;
import java.util.Random;
import java.util.concurrent.*;
import static com.google.common.base.Preconditions.*;
@ -48,6 +49,7 @@ class MetricsSinkAdapter implements SinkQueue.Consumer<MetricsBuffer> {
private volatile boolean stopping = false;
private volatile boolean inError = false;
private final int period, firstRetryDelay, retryCount;
private final long oobPutTimeout;
private final float retryBackoff;
private final MetricsRegistry registry = new MetricsRegistry("sinkadapter");
private final MutableStat latency;
@ -69,6 +71,8 @@ class MetricsSinkAdapter implements SinkQueue.Consumer<MetricsBuffer> {
this.period = checkArg(period, period > 0, "period");
firstRetryDelay = checkArg(retryDelay, retryDelay > 0, "retry delay");
this.retryBackoff = checkArg(retryBackoff, retryBackoff>1, "retry backoff");
oobPutTimeout = (long)
(firstRetryDelay * Math.pow(retryBackoff, retryCount) * 1000);
this.retryCount = retryCount;
this.queue = new SinkQueue<MetricsBuffer>(checkArg(queueCapacity,
queueCapacity > 0, "queue capacity"));
@ -96,6 +100,23 @@ class MetricsSinkAdapter implements SinkQueue.Consumer<MetricsBuffer> {
return true; // OK
}
public boolean putMetricsImmediate(MetricsBuffer buffer) {
WaitableMetricsBuffer waitableBuffer =
new WaitableMetricsBuffer(buffer);
if (!queue.enqueue(waitableBuffer)) {
LOG.warn(name + " has a full queue and can't consume the given metrics.");
dropped.incr();
return false;
}
if (!waitableBuffer.waitTillNotified(oobPutTimeout)) {
LOG.warn(name +
" couldn't fulfill an immediate putMetrics request in time." +
" Abandoning.");
return false;
}
return true;
}
void publishMetricsFromQueue() {
int retryDelay = firstRetryDelay;
int n = retryCount;
@ -158,6 +179,9 @@ class MetricsSinkAdapter implements SinkQueue.Consumer<MetricsBuffer> {
sink.flush();
latency.add(Time.now() - ts);
}
if (buffer instanceof WaitableMetricsBuffer) {
((WaitableMetricsBuffer)buffer).notifyAnyWaiters();
}
LOG.debug("Done");
}
@ -191,4 +215,26 @@ class MetricsSinkAdapter implements SinkQueue.Consumer<MetricsBuffer> {
MetricsSink sink() {
return sink;
}
static class WaitableMetricsBuffer extends MetricsBuffer {
private final Semaphore notificationSemaphore =
new Semaphore(0);
public WaitableMetricsBuffer(MetricsBuffer metricsBuffer) {
super(metricsBuffer);
}
public boolean waitTillNotified(long millisecondsToWait) {
try {
return notificationSemaphore.tryAcquire(millisecondsToWait,
TimeUnit.MILLISECONDS);
} catch (InterruptedException e) {
return false;
}
}
public void notifyAnyWaiters() {
notificationSemaphore.release();
}
}
}

24
hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/metrics2/impl/MetricsSystemImpl.java
View File

@ -344,7 +344,17 @@ public class MetricsSystemImpl extends MetricsSystem implements MetricsSource {
synchronized void onTimerEvent() {
logicalTime += period;
if (sinks.size() > 0) {
publishMetrics(sampleMetrics());
publishMetrics(sampleMetrics(), false);
}
}
/**
* Requests an immediate publish of all metrics from sources to sinks.
*/
@Override
public void publishMetricsNow() {
if (sinks.size() > 0) {
publishMetrics(sampleMetrics(), true);
}
}
@ -380,12 +390,20 @@ public class MetricsSystemImpl extends MetricsSystem implements MetricsSource {
/**
* Publish a metrics snapshot to all the sinks
* @param buffer the metrics snapshot to publish
* @param immediate indicates that we should publish metrics immediately
* instead of using a separate thread.
*/
synchronized void publishMetrics(MetricsBuffer buffer) {
synchronized void publishMetrics(MetricsBuffer buffer, boolean immediate) {
int dropped = 0;
for (MetricsSinkAdapter sa : sinks.values()) {
long startTime = Time.now();
dropped += sa.putMetrics(buffer, logicalTime) ? 0 : 1;
boolean result;
if (immediate) {
result = sa.putMetricsImmediate(buffer);
} else {
result = sa.putMetrics(buffer, logicalTime);
}
dropped += result ? 0 : 1;
publishStat.add(Time.now() - startTime);
}
droppedPubAll.incr(dropped);

29
hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/metrics2/impl/TestGangliaMetrics.java
View File

@ -29,8 +29,6 @@ import java.util.ArrayList;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.TimeUnit;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
@ -115,31 +113,23 @@ public class TestGangliaMetrics {
final int expectedCountFromGanglia30 = expectedMetrics.length;
final int expectedCountFromGanglia31 = 2 * expectedMetrics.length;
// use latch to make sure we received required records before shutting
// down the MetricSystem
CountDownLatch latch = new CountDownLatch(
expectedCountFromGanglia30 + expectedCountFromGanglia31);
// Setup test for GangliaSink30
AbstractGangliaSink gsink30 = new GangliaSink30();
gsink30.init(cb.subset("test"));
MockDatagramSocket mockds30 = new MockDatagramSocket(latch);
MockDatagramSocket mockds30 = new MockDatagramSocket();
GangliaMetricsTestHelper.setDatagramSocket(gsink30, mockds30);
// Setup test for GangliaSink31
AbstractGangliaSink gsink31 = new GangliaSink31();
gsink31.init(cb.subset("test"));
MockDatagramSocket mockds31 = new MockDatagramSocket(latch);
MockDatagramSocket mockds31 = new MockDatagramSocket();
GangliaMetricsTestHelper.setDatagramSocket(gsink31, mockds31);
// register the sinks
ms.register("gsink30", "gsink30 desc", gsink30);
ms.register("gsink31", "gsink31 desc", gsink31);
ms.onTimerEvent(); // trigger something interesting
ms.publishMetricsNow(); // publish the metrics
// wait for all records and the stop MetricSystem. Without this
// sometime the ms gets shutdown before all the sinks have consumed
latch.await(200, TimeUnit.MILLISECONDS);
ms.stop();
// check GanfliaSink30 data
@ -198,7 +188,6 @@ public class TestGangliaMetrics {
*/
private class MockDatagramSocket extends DatagramSocket {
private ArrayList<byte[]> capture;
private CountDownLatch latch;
/**
* @throws SocketException
@ -207,15 +196,6 @@ public class TestGangliaMetrics {
capture = new ArrayList<byte[]>();
}
/**
* @param latch
* @throws SocketException
*/
public MockDatagramSocket(CountDownLatch latch) throws SocketException {
this();
this.latch = latch;
}
/* (non-Javadoc)
* @see java.net.DatagramSocket#send(java.net.DatagramPacket)
*/
@ -225,9 +205,6 @@ public class TestGangliaMetrics {
byte[] bytes = new byte[p.getLength()];
System.arraycopy(p.getData(), p.getOffset(), bytes, 0, p.getLength());
capture.add(bytes);
// decrement the latch
latch.countDown();
}
/**

188
hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/metrics2/impl/TestMetricsSystemImpl.java
View File

@ -18,7 +18,11 @@
package org.apache.hadoop.metrics2.impl;
import java.util.List;
import java.util.*;
import java.util.concurrent.*;
import java.util.concurrent.atomic.*;
import javax.annotation.Nullable;
import org.junit.Test;
import org.junit.runner.RunWith;
@ -26,9 +30,11 @@ import org.junit.runner.RunWith;
import org.mockito.ArgumentCaptor;
import org.mockito.Captor;
import org.mockito.runners.MockitoJUnitRunner;
import static org.junit.Assert.*;
import static org.mockito.Mockito.*;
import com.google.common.base.Predicate;
import com.google.common.collect.Iterables;
import org.apache.commons.configuration.SubsetConfiguration;
@ -36,6 +42,8 @@ import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.metrics2.MetricsException;
import static org.apache.hadoop.test.MoreAsserts.*;
import org.apache.hadoop.metrics2.AbstractMetric;
import org.apache.hadoop.metrics2.MetricsRecord;
import org.apache.hadoop.metrics2.MetricsSink;
import org.apache.hadoop.metrics2.MetricsSource;
@ -47,6 +55,7 @@ import org.apache.hadoop.metrics2.lib.MetricsRegistry;
import org.apache.hadoop.metrics2.lib.MutableCounterLong;
import org.apache.hadoop.metrics2.lib.MutableRate;
import org.apache.hadoop.metrics2.lib.MutableGaugeLong;
import org.apache.hadoop.util.StringUtils;
/**
* Test the MetricsSystemImpl class
@ -72,7 +81,7 @@ public class TestMetricsSystemImpl {
}
@Test public void testInitFirst() throws Exception {
ConfigBuilder cb = new ConfigBuilder().add("*.period", 8)
new ConfigBuilder().add("*.period", 8)
//.add("test.sink.plugin.urls", getPluginUrlsAsString())
.add("test.sink.test.class", TestSink.class.getName())
.add("test.*.source.filter.exclude", "s0")
@ -93,8 +102,9 @@ public class TestMetricsSystemImpl {
MetricsSink sink2 = mock(MetricsSink.class);
ms.registerSink("sink1", "sink1 desc", sink1);
ms.registerSink("sink2", "sink2 desc", sink2);
ms.onTimerEvent(); // trigger something interesting
ms.publishMetricsNow(); // publish the metrics
ms.stop();
ms.shutdown();
verify(sink1, times(2)).putMetrics(r1.capture());
List<MetricsRecord> mr1 = r1.getAllValues();
@ -104,6 +114,177 @@ public class TestMetricsSystemImpl {
assertEquals("output", mr1, mr2);
}
@Test public void testMultiThreadedPublish() throws Exception {
new ConfigBuilder().add("*.period", 80)
.add("test.sink.Collector.queue.capacity", "20")
.save(TestMetricsConfig.getTestFilename("hadoop-metrics2-test"));
final MetricsSystemImpl ms = new MetricsSystemImpl("Test");
ms.start();
final int numThreads = 10;
final CollectingSink sink = new CollectingSink(numThreads);
ms.registerSink("Collector",
"Collector of values from all threads.", sink);
final TestSource[] sources = new TestSource[numThreads];
final Thread[] threads = new Thread[numThreads];
final String[] results = new String[numThreads];
final CyclicBarrier barrier1 = new CyclicBarrier(numThreads),
barrier2 = new CyclicBarrier(numThreads);
for (int i = 0; i < numThreads; i++) {
sources[i] = ms.register("threadSource" + i,
"A source of my threaded goodness.",
new TestSource("threadSourceRec" + i));
threads[i] = new Thread(new Runnable() {
private boolean safeAwait(int mySource, CyclicBarrier barrier) {
try {
barrier1.await(2, TimeUnit.SECONDS);
} catch (InterruptedException e) {
results[mySource] = "Interrupted";
return false;
} catch (BrokenBarrierException e) {
results[mySource] = "Broken Barrier";
return false;
} catch (TimeoutException e) {
results[mySource] = "Timed out on barrier";
return false;
}
return true;
}
@Override
public void run() {
int mySource = Integer.parseInt(Thread.currentThread().getName());
if (sink.collected[mySource].get() != 0L) {
results[mySource] = "Someone else collected my metric!";
return;
}
// Wait for all the threads to come here so we can hammer
// the system at the same time
if (!safeAwait(mySource, barrier1)) return;
sources[mySource].g1.set(230);
ms.publishMetricsNow();
// Since some other thread may have snatched my metric,
// I need to wait for the threads to finish before checking.
if (!safeAwait(mySource, barrier2)) return;
if (sink.collected[mySource].get() != 230L) {
results[mySource] = "Metric not collected!";
return;
}
results[mySource] = "Passed";
}
}, "" + i);
}
for (Thread t : threads)
t.start();
for (Thread t : threads)
t.join();
assertEquals(0L, ms.droppedPubAll.value());
assertTrue(StringUtils.join("\n", Arrays.asList(results)),
Iterables.all(Arrays.asList(results), new Predicate<String>() {
@Override
public boolean apply(@Nullable String input) {
return input.equalsIgnoreCase("Passed");
}
}));
ms.stop();
ms.shutdown();
}
private static class CollectingSink implements MetricsSink {
private final AtomicLong[] collected;
public CollectingSink(int capacity) {
collected = new AtomicLong[capacity];
for (int i = 0; i < capacity; i++) {
collected[i] = new AtomicLong();
}
}
@Override
public void init(SubsetConfiguration conf) {
}
@Override
public void putMetrics(MetricsRecord record) {
final String prefix = "threadSourceRec";
if (record.name().startsWith(prefix)) {
final int recordNumber = Integer.parseInt(
record.name().substring(prefix.length()));
ArrayList<String> names = new ArrayList<String>();
for (AbstractMetric m : record.metrics()) {
if (m.name().equalsIgnoreCase("g1")) {
collected[recordNumber].set(m.value().longValue());
return;
}
names.add(m.name());
}
}
}
@Override
public void flush() {
}
}
@Test public void testHangingSink() {
new ConfigBuilder().add("*.period", 8)
.add("test.sink.test.class", TestSink.class.getName())
.add("test.sink.hanging.retry.delay", "1")
.add("test.sink.hanging.retry.backoff", "1.01")
.add("test.sink.hanging.retry.count", "0")
.save(TestMetricsConfig.getTestFilename("hadoop-metrics2-test"));
MetricsSystemImpl ms = new MetricsSystemImpl("Test");
ms.start();
TestSource s = ms.register("s3", "s3 desc", new TestSource("s3rec"));
s.c1.incr();
HangingSink hanging = new HangingSink();
ms.registerSink("hanging", "Hang the sink!", hanging);
ms.publishMetricsNow();
assertEquals(1L, ms.droppedPubAll.value());
assertFalse(hanging.getInterrupted());
ms.stop();
ms.shutdown();
assertTrue(hanging.getInterrupted());
assertTrue("The sink didn't get called after its first hang " +
"for subsequent records.", hanging.getGotCalledSecondTime());
}
private static class HangingSink implements MetricsSink {
private volatile boolean interrupted;
private boolean gotCalledSecondTime;
private boolean firstTime = true;
public boolean getGotCalledSecondTime() {
return gotCalledSecondTime;
}
public boolean getInterrupted() {
return interrupted;
}
@Override
public void init(SubsetConfiguration conf) {
}
@Override
public void putMetrics(MetricsRecord record) {
// No need to hang every time, just the first record.
if (!firstTime) {
gotCalledSecondTime = true;
return;
}
firstTime = false;
try {
Thread.sleep(10 * 1000);
} catch (InterruptedException ex) {
interrupted = true;
}
}
@Override
public void flush() {
}
}
@Test public void testRegisterDups() {
MetricsSystem ms = new MetricsSystemImpl();
TestSource ts1 = new TestSource("ts1");
@ -116,6 +297,7 @@ public class TestMetricsSystemImpl {
MetricsSource s2 = ms.getSource("ts1");
assertNotNull(s2);
assertNotSame(s1, s2);
ms.shutdown();
}
@Test(expected=MetricsException.class) public void testRegisterDupError() {

10
hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
View File

@ -176,6 +176,9 @@ Trunk (Unreleased)
HDFS-4209. Clean up the addNode/addChild/addChildNoQuotaCheck methods in
FSDirectory and INodeDirectory. (szetszwo)
HDFS-3358. Specify explicitly that the NN UI status total is talking
of persistent objects on heap. (harsh)
OPTIMIZATIONS
BUG FIXES
@ -646,6 +649,8 @@ Release 2.0.3-alpha - Unreleased
of it is undefined after the iteration or modifications of the map.
(szetszwo)
HDFS-4231. BackupNode: Introduce BackupState. (shv)
Release 2.0.2-alpha - 2012-09-07
INCOMPATIBLE CHANGES
@ -2035,6 +2040,11 @@ Release 0.23.6 - UNRELEASED
BUG FIXES
HDFS-4247. saveNamespace should be tolerant of dangling lease (daryn)
HDFS-4248. Renaming directories may incorrectly remove the paths in leases
under the tree. (daryn via szetszwo)
Release 0.23.5 - UNRELEASED
INCOMPATIBLE CHANGES

49
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/BackupNode.java
View File

@ -24,6 +24,7 @@ import java.net.SocketTimeoutException;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.CommonConfigurationKeys;
import org.apache.hadoop.ha.ServiceFailedException;
import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.hdfs.DFSUtil;
import org.apache.hadoop.hdfs.NameNodeProxies;
@ -35,6 +36,7 @@ import org.apache.hadoop.hdfs.protocolPB.JournalProtocolPB;
import org.apache.hadoop.hdfs.protocolPB.JournalProtocolServerSideTranslatorPB;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.NamenodeRole;
import org.apache.hadoop.hdfs.server.common.Storage;
import org.apache.hadoop.hdfs.server.namenode.ha.HAState;
import org.apache.hadoop.hdfs.server.protocol.FenceResponse;
import org.apache.hadoop.hdfs.server.protocol.JournalInfo;
import org.apache.hadoop.hdfs.server.protocol.JournalProtocol;
@ -416,12 +418,21 @@ public class BackupNode extends NameNode {
}
@Override
protected String getNameServiceId(Configuration conf) {
return DFSUtil.getBackupNameServiceId(conf);
}
protected HAState createHAState() {
return new BackupState();
}
@Override // NameNode
protected NameNodeHAContext createHAContext() {
return new BNHAContext();
}
private class BNHAContext extends NameNodeHAContext {
@Override // NameNode
@Override // NameNodeHAContext
public void checkOperation(OperationCategory op)
throws StandbyException {
if (op == OperationCategory.UNCHECKED ||
@ -435,10 +446,42 @@ public class BackupNode extends NameNode {
throw new StandbyException(msg);
}
}
@Override // NameNodeHAContext
public void prepareToStopStandbyServices() throws ServiceFailedException {
}
/**
* Start services for BackupNode.
* <p>
* The following services should be muted
* (not run or not pass any control commands to DataNodes)
* on BackupNode:
* {@link LeaseManager.Monitor} protected by SafeMode.
* {@link BlockManager.ReplicationMonitor} protected by SafeMode.
* {@link HeartbeatManager.Monitor} protected by SafeMode.
* {@link DecommissionManager.Monitor} need to prohibit refreshNodes().
* {@link PendingReplicationBlocks.PendingReplicationMonitor} harmless,
* because ReplicationMonitor is muted.
*/
@Override
public void startActiveServices() throws IOException {
try {
namesystem.startActiveServices();
} catch (Throwable t) {
doImmediateShutdown(t);
}
}
@Override
protected String getNameServiceId(Configuration conf) {
return DFSUtil.getBackupNameServiceId(conf);
public void stopActiveServices() throws IOException {
try {
if (namesystem != null) {
namesystem.stopActiveServices();
}
} catch (Throwable t) {
doImmediateShutdown(t);
}
}
}
}

51
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/BackupState.java Normal file
View File

@ -0,0 +1,51 @@
package org.apache.hadoop.hdfs.server.namenode;
import java.io.IOException;
import org.apache.hadoop.ha.HAServiceProtocol.HAServiceState;
import org.apache.hadoop.ha.ServiceFailedException;
import org.apache.hadoop.hdfs.server.namenode.NameNode.OperationCategory;
import org.apache.hadoop.hdfs.server.namenode.ha.HAContext;
import org.apache.hadoop.hdfs.server.namenode.ha.HAState;
import org.apache.hadoop.ipc.StandbyException;
public class BackupState extends HAState {
public BackupState() {
super(HAServiceState.STANDBY);
}
@Override // HAState
public void checkOperation(HAContext context, OperationCategory op)
throws StandbyException {
context.checkOperation(op);
}
@Override // HAState
public boolean shouldPopulateReplQueues() {
return false;
}
@Override
public void enterState(HAContext context) throws ServiceFailedException {
try {
context.startActiveServices();
} catch (IOException e) {
throw new ServiceFailedException("Failed to start backup services", e);
}
}
@Override
public void exitState(HAContext context) throws ServiceFailedException {
try {
context.stopActiveServices();
} catch (IOException e) {
throw new ServiceFailedException("Failed to stop backup services", e);
}
}
@Override
public void prepareToExitState(HAContext context) throws ServiceFailedException {
context.prepareToStopStandbyServices();
}
}

73
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSDirectory.java
View File

@ -383,7 +383,7 @@ public class FSDirectory implements Closeable {
writeLock();
try {
// file is closed
file.setModificationTimeForce(now);
file.setModificationTime(now);
fsImage.getEditLog().logCloseFile(path, file);
if (NameNode.stateChangeLog.isDebugEnabled()) {
NameNode.stateChangeLog.debug("DIR* FSDirectory.closeFile: "
@ -585,8 +585,10 @@ public class FSDirectory implements Closeable {
+ src + " is renamed to " + dst);
}
// update modification time of dst and the parent of src
srcInodes[srcInodes.length-2].setModificationTime(timestamp);
dstInodes[dstInodes.length-2].setModificationTime(timestamp);
srcInodes[srcInodes.length-2].updateModificationTime(timestamp);
dstInodes[dstInodes.length-2].updateModificationTime(timestamp);
// update moved leases with new filename
getFSNamesystem().unprotectedChangeLease(src, dst);
return true;
}
} finally {
@ -750,8 +752,10 @@ public class FSDirectory implements Closeable {
"DIR* FSDirectory.unprotectedRenameTo: " + src
+ " is renamed to " + dst);
}
srcInodes[srcInodes.length - 2].setModificationTime(timestamp);
dstInodes[dstInodes.length - 2].setModificationTime(timestamp);
srcInodes[srcInodes.length - 2].updateModificationTime(timestamp);
dstInodes[dstInodes.length - 2].updateModificationTime(timestamp);
// update moved lease with new filename
getFSNamesystem().unprotectedChangeLease(src, dst);
// Collect the blocks and remove the lease for previous dst
int filesDeleted = 0;
@ -986,12 +990,12 @@ public class FSDirectory implements Closeable {
if(nodeToRemove == null) continue;
nodeToRemove.setBlocks(null);
trgParent.removeChild(nodeToRemove);
trgParent.removeChild(nodeToRemove, trgINodesInPath.getLatestSnapshot());
count++;
}
trgInode.setModificationTimeForce(timestamp);
trgParent.setModificationTime(timestamp);
trgInode.setModificationTime(timestamp);
trgParent.updateModificationTime(timestamp);
// update quota on the parent directory ('count' files removed, 0 space)
unprotectedUpdateCount(trgINodesInPath, trgINodes.length-1, -count, 0);
}
@ -1129,7 +1133,7 @@ public class FSDirectory implements Closeable {
return 0;
}
// set the parent's modification time
inodes[inodes.length - 2].setModificationTime(mtime);
inodes[inodes.length - 2].updateModificationTime(mtime);
int filesRemoved = targetNode.collectSubtreeBlocksAndClear(collectedBlocks);
if (NameNode.stateChangeLog.isDebugEnabled()) {
NameNode.stateChangeLog.debug("DIR* FSDirectory.unprotectedDelete: "
@ -1165,10 +1169,10 @@ public class FSDirectory implements Closeable {
/**
* Replaces the specified INode.
*/
private void replaceINodeUnsynced(String path, INode oldnode, INode newnode
) throws IOException {
private void replaceINodeUnsynced(String path, INode oldnode, INode newnode,
Snapshot latestSnapshot) throws IOException {
//remove the old node from the namespace
if (!oldnode.removeNode()) {
if (!oldnode.removeNode(latestSnapshot)) {
final String mess = "FSDirectory.replaceINodeUnsynced: failed to remove "
+ path;
NameNode.stateChangeLog.warn("DIR* " + mess);
@ -1183,10 +1187,10 @@ public class FSDirectory implements Closeable {
* Replaces the specified INodeDirectory.
*/
public void replaceINodeDirectory(String path, INodeDirectory oldnode,
INodeDirectory newnode) throws IOException {
INodeDirectory newnode, Snapshot latestSnapshot) throws IOException {
writeLock();
try {
replaceINodeUnsynced(path, oldnode, newnode);
replaceINodeUnsynced(path, oldnode, newnode, latestSnapshot);
//update children's parent directory
for(INode i : newnode.getChildrenList(null)) {
@ -1200,11 +1204,11 @@ public class FSDirectory implements Closeable {
/**
* Replaces the specified INodeFile with the specified one.
*/
public void replaceNode(String path, INodeFile oldnode, INodeFile newnode
) throws IOException {
public void replaceNode(String path, INodeFile oldnode, INodeFile newnode,
Snapshot latestSnapshot) throws IOException {
writeLock();
try {
replaceINodeUnsynced(path, oldnode, newnode);
replaceINodeUnsynced(path, oldnode, newnode, latestSnapshot);
//Currently, oldnode and newnode are assumed to contain the same blocks.
//Otherwise, blocks need to be removed from the blocksMap.
@ -1273,13 +1277,9 @@ public class FSDirectory implements Closeable {
String srcs = normalizePath(src);
readLock();
try {
INode targetNode = rootDir.getNode(srcs, resolveLink);
if (targetNode == null) {
return null;
}
else {
return createFileStatus(HdfsFileStatus.EMPTY_NAME, targetNode);
}
final INodesInPath inodesInPath = rootDir.getINodesInPath(srcs, resolveLink);
final INode i = inodesInPath.getINode(0);
return i == null? null: createFileStatus(HdfsFileStatus.EMPTY_NAME, i);
} finally {
readUnlock();
}
@ -1303,9 +1303,16 @@ public class FSDirectory implements Closeable {
* Get {@link INode} associated with the file / directory.
*/
public INode getINode(String src) throws UnresolvedLinkException {
return getINodesInPath(src).getINode(0);
}
/**
* Get {@link INode} associated with the file / directory.
*/
public INodesInPath getINodesInPath(String src) throws UnresolvedLinkException {
readLock();
try {
return rootDir.getNode(src, true);
return rootDir.getINodesInPath(src, true);
} finally {
readUnlock();
}
@ -1800,7 +1807,8 @@ public class FSDirectory implements Closeable {
if (inodes[pos-1] == null) {
throw new NullPointerException("Panic: parent does not exist");
}
final boolean added = ((INodeDirectory)inodes[pos-1]).addChild(child, true);
final boolean added = ((INodeDirectory)inodes[pos-1]).addChild(child, true,
inodesInPath.getLatestSnapshot());
if (!added) {
updateCount(inodesInPath, pos, -counts.getNsCount(), -counts.getDsCount(), true);
}
@ -1824,7 +1832,8 @@ public class FSDirectory implements Closeable {
private INode removeLastINode(final INodesInPath inodesInPath) {
final INode[] inodes = inodesInPath.getINodes();
final int pos = inodes.length - 1;
INode removedNode = ((INodeDirectory)inodes[pos-1]).removeChild(inodes[pos]);
INode removedNode = ((INodeDirectory)inodes[pos-1]).removeChild(inodes[pos],
inodesInPath.getLatestSnapshot());
if (removedNode != null) {
INode.DirCounts counts = new INode.DirCounts();
removedNode.spaceConsumedInTree(counts);
@ -1965,8 +1974,8 @@ public class FSDirectory implements Closeable {
}
String srcs = normalizePath(src);
final INode[] inodes = rootDir.getMutableINodesInPath(srcs, true)
.getINodes();
final INodesInPath inodesInPath = rootDir.getMutableINodesInPath(srcs, true);
final INode[] inodes = inodesInPath.getINodes();
INodeDirectory dirNode = INodeDirectory.valueOf(inodes[inodes.length-1], srcs);
if (dirNode.isRoot() && nsQuota == HdfsConstants.QUOTA_RESET) {
throw new IllegalArgumentException("Cannot clear namespace quota on root.");
@ -1988,7 +1997,7 @@ public class FSDirectory implements Closeable {
INodeDirectory newNode = new INodeDirectory(dirNode);
INodeDirectory parent = (INodeDirectory)inodes[inodes.length-2];
dirNode = newNode;
parent.replaceChild(newNode);
parent.replaceChild(newNode, inodesInPath.getLatestSnapshot());
}
} else {
// a non-quota directory; so replace it with a directory with quota
@ -1997,7 +2006,7 @@ public class FSDirectory implements Closeable {
// non-root directory node; parent != null
INodeDirectory parent = (INodeDirectory)inodes[inodes.length-2];
dirNode = newNode;
parent.replaceChild(newNode);
parent.replaceChild(newNode, inodesInPath.getLatestSnapshot());
}
return (oldNsQuota != nsQuota || oldDsQuota != dsQuota) ? dirNode : null;
}
@ -2061,7 +2070,7 @@ public class FSDirectory implements Closeable {
assert hasWriteLock();
boolean status = false;
if (mtime != -1) {
inode.setModificationTimeForce(mtime);
inode.setModificationTime(mtime);
status = true;
}
if (atime != -1) {

28
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSEditLogLoader.java
View File

@ -31,7 +31,6 @@ import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
import org.apache.hadoop.hdfs.protocol.LayoutVersion;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoUnderConstruction;
@ -58,6 +57,7 @@ import org.apache.hadoop.hdfs.server.namenode.FSEditLogOp.SymlinkOp;
import org.apache.hadoop.hdfs.server.namenode.FSEditLogOp.TimesOp;
import org.apache.hadoop.hdfs.server.namenode.FSEditLogOp.UpdateBlocksOp;
import org.apache.hadoop.hdfs.server.namenode.FSEditLogOp.UpdateMasterKeyOp;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory.INodesInPath;
import org.apache.hadoop.hdfs.server.namenode.LeaseManager.Lease;
import org.apache.hadoop.hdfs.util.Holder;
@ -246,7 +246,8 @@ public class FSEditLogLoader {
// 3. OP_ADD to open file for append
// See if the file already exists (persistBlocks call)
INodeFile oldFile = getINodeFile(fsDir, addCloseOp.path);
final INodesInPath inodesInPath = fsDir.getINodesInPath(addCloseOp.path);
INodeFile oldFile = toINodeFile(inodesInPath.getINode(0), addCloseOp.path);
INodeFile newFile = oldFile;
if (oldFile == null) { // this is OP_ADD on a new file (case 1)
// versions > 0 support per file replication
@ -272,7 +273,7 @@ public class FSEditLogLoader {
}
fsNamesys.prepareFileForWrite(addCloseOp.path, oldFile,
addCloseOp.clientName, addCloseOp.clientMachine, null,
false);
false, inodesInPath.getLatestSnapshot());
newFile = getINodeFile(fsDir, addCloseOp.path);
}
}
@ -282,7 +283,7 @@ public class FSEditLogLoader {
// Update the salient file attributes.
newFile.setAccessTime(addCloseOp.atime);
newFile.setModificationTimeForce(addCloseOp.mtime);
newFile.setModificationTime(addCloseOp.mtime);
updateBlocks(fsDir, addCloseOp, newFile);
break;
}
@ -296,7 +297,8 @@ public class FSEditLogLoader {
" clientMachine " + addCloseOp.clientMachine);
}
INodeFile oldFile = getINodeFile(fsDir, addCloseOp.path);
final INodesInPath inodesInPath = fsDir.getINodesInPath(addCloseOp.path);
INodeFile oldFile = toINodeFile(inodesInPath.getINode(0), addCloseOp.path);
if (oldFile == null) {
throw new IOException("Operation trying to close non-existent file " +
addCloseOp.path);
@ -304,7 +306,7 @@ public class FSEditLogLoader {
// Update the salient file attributes.
oldFile.setAccessTime(addCloseOp.atime);
oldFile.setModificationTimeForce(addCloseOp.mtime);
oldFile.setModificationTime(addCloseOp.mtime);
updateBlocks(fsDir, addCloseOp, oldFile);
// Now close the file
@ -322,7 +324,8 @@ public class FSEditLogLoader {
INodeFileUnderConstruction ucFile = (INodeFileUnderConstruction) oldFile;
fsNamesys.leaseManager.removeLeaseWithPrefixPath(addCloseOp.path);
INodeFile newFile = ucFile.convertToInodeFile();
fsDir.replaceNode(addCloseOp.path, ucFile, newFile);
fsDir.replaceNode(addCloseOp.path, ucFile, newFile,
inodesInPath.getLatestSnapshot());
}
break;
}
@ -360,10 +363,8 @@ public class FSEditLogLoader {
}
case OP_RENAME_OLD: {
RenameOldOp renameOp = (RenameOldOp)op;
HdfsFileStatus dinfo = fsDir.getFileInfo(renameOp.dst, false);
fsDir.unprotectedRenameTo(renameOp.src, renameOp.dst,
renameOp.timestamp);
fsNamesys.unprotectedChangeLease(renameOp.src, renameOp.dst, dinfo);
break;
}
case OP_DELETE: {
@ -433,11 +434,8 @@ public class FSEditLogLoader {
}
case OP_RENAME: {
RenameOp renameOp = (RenameOp)op;
HdfsFileStatus dinfo = fsDir.getFileInfo(renameOp.dst, false);
fsDir.unprotectedRenameTo(renameOp.src, renameOp.dst,
renameOp.timestamp, renameOp.options);
fsNamesys.unprotectedChangeLease(renameOp.src, renameOp.dst, dinfo);
break;
}
case OP_GET_DELEGATION_TOKEN: {
@ -512,7 +510,11 @@ public class FSEditLogLoader {
private static INodeFile getINodeFile(FSDirectory fsDir, String path)
throws IOException {
INode inode = fsDir.getINode(path);
return toINodeFile(fsDir.getINode(path), path);
}
private static INodeFile toINodeFile(INode inode, String path)
throws IOException {
if (inode != null) {
if (!(inode instanceof INodeFile)) {
throw new IOException("Operation trying to get non-file " + path);

10
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSImageFormat.java
View File

@ -44,6 +44,7 @@ import org.apache.hadoop.hdfs.protocol.LayoutVersion.Feature;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
import org.apache.hadoop.hdfs.server.common.InconsistentFSStateException;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory.INodesInPath;
import org.apache.hadoop.hdfs.util.ReadOnlyList;
import org.apache.hadoop.io.MD5Hash;
import org.apache.hadoop.io.Text;
@ -202,7 +203,7 @@ class FSImageFormat {
if (nsQuota != -1 || dsQuota != -1) {
fsDir.rootDir.setQuota(nsQuota, dsQuota);
}
fsDir.rootDir.setModificationTime(root.getModificationTime());
fsDir.rootDir.cloneModificationTime(root);
fsDir.rootDir.clonePermissionStatus(root);
}
@ -305,7 +306,7 @@ class FSImageFormat {
*/
void addToParent(INodeDirectory parent, INode child) {
// NOTE: This does not update space counts for parents
if (!parent.addChild(child, false)) {
if (!parent.addChild(child, false, null)) {
return;
}
namesystem.dir.cacheName(child);
@ -388,8 +389,9 @@ class FSImageFormat {
// verify that file exists in namespace
String path = cons.getLocalName();
INodeFile oldnode = INodeFile.valueOf(fsDir.getINode(path), path);
fsDir.replaceNode(path, oldnode, cons);
final INodesInPath inodesInPath = fsDir.getINodesInPath(path);
INodeFile oldnode = INodeFile.valueOf(inodesInPath.getINode(0), path);
fsDir.replaceNode(path, oldnode, cons, inodesInPath.getLatestSnapshot());
namesystem.leaseManager.addLease(cons.getClientName(), path);
}
}

121
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSNamesystem.java
View File

@ -121,6 +121,7 @@ import org.apache.hadoop.fs.UnresolvedLinkException;
import org.apache.hadoop.fs.permission.FsAction;
import org.apache.hadoop.fs.permission.FsPermission;
import org.apache.hadoop.fs.permission.PermissionStatus;
import org.apache.hadoop.ha.HAServiceProtocol.HAServiceState;
import org.apache.hadoop.ha.ServiceFailedException;
import org.apache.hadoop.hdfs.DFSUtil;
import org.apache.hadoop.hdfs.HAUtil;
@ -165,7 +166,6 @@ import org.apache.hadoop.hdfs.server.namenode.INode.BlocksMapUpdateInfo;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory.INodesInPath;
import org.apache.hadoop.hdfs.server.namenode.LeaseManager.Lease;
import org.apache.hadoop.hdfs.server.namenode.NameNode.OperationCategory;
import org.apache.hadoop.hdfs.server.namenode.ha.ActiveState;
import org.apache.hadoop.hdfs.server.namenode.ha.EditLogTailer;
import org.apache.hadoop.hdfs.server.namenode.ha.HAContext;
import org.apache.hadoop.hdfs.server.namenode.ha.HAState;
@ -173,7 +173,9 @@ import org.apache.hadoop.hdfs.server.namenode.ha.StandbyCheckpointer;
import org.apache.hadoop.hdfs.server.namenode.ha.StandbyState;
import org.apache.hadoop.hdfs.server.namenode.metrics.FSNamesystemMBean;
import org.apache.hadoop.hdfs.server.namenode.metrics.NameNodeMetrics;
import org.apache.hadoop.hdfs.server.namenode.snapshot.INodeDirectorySnapshottable;
import org.apache.hadoop.hdfs.server.namenode.snapshot.INodeFileWithLink;
import org.apache.hadoop.hdfs.server.namenode.snapshot.Snapshot;
import org.apache.hadoop.hdfs.server.namenode.snapshot.SnapshotManager;
import org.apache.hadoop.hdfs.server.namenode.web.resources.NamenodeWebHdfsMethods;
import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand;
@ -1036,7 +1038,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
long totalInodes = this.dir.totalInodes();
long totalBlocks = this.getBlocksTotal();
out.println(totalInodes + " files and directories, " + totalBlocks
+ " blocks = " + (totalInodes + totalBlocks) + " total");
+ " blocks = " + (totalInodes + totalBlocks)
+ " total filesystem objects");
blockManager.metaSave(out);
}
@ -1820,7 +1823,9 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
try {
blockManager.verifyReplication(src, replication, clientMachine);
boolean create = flag.contains(CreateFlag.CREATE);
final INode myFile = dir.getINode(src);
final INodesInPath inodesInPath = dir.getINodesInPath(src);
final INode myFile = inodesInPath.getINode(0);
if (myFile == null) {
if (!create) {
throw new FileNotFoundException("failed to overwrite or append to non-existent file "
@ -1847,8 +1852,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
if (append && myFile != null) {
final INodeFile f = INodeFile.valueOf(myFile, src);
return prepareFileForWrite(
src, f, holder, clientMachine, clientNode, true);
return prepareFileForWrite(src, f, holder, clientMachine, clientNode,
true, inodesInPath.getLatestSnapshot());
} else {
// Now we can add the name to the filesystem. This file has no
// blocks associated with it.
@ -1896,7 +1901,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
*/
LocatedBlock prepareFileForWrite(String src, INodeFile file,
String leaseHolder, String clientMachine, DatanodeDescriptor clientNode,
boolean writeToEditLog) throws IOException {
boolean writeToEditLog, Snapshot latestSnapshot) throws IOException {
//TODO SNAPSHOT: INodeFileUnderConstruction with link
INodeFileUnderConstruction cons = new INodeFileUnderConstruction(
file.getLocalNameBytes(),
@ -1908,7 +1913,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
leaseHolder,
clientMachine,
clientNode);
dir.replaceNode(src, file, cons);
dir.replaceNode(src, file, cons, latestSnapshot);
leaseManager.addLease(cons.getClientName(), src);
LocatedBlock ret = blockManager.convertLastBlockToUnderConstruction(cons);
@ -2158,7 +2163,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// have we exceeded the configured limit of fs objects.
checkFsObjectLimit();
INodeFileUnderConstruction pendingFile = checkLease(src, clientName);
final INodeFileUnderConstruction pendingFile = checkLease(
src, clientName, dir.getINode(src));
BlockInfo lastBlockInFile = pendingFile.getLastBlock();
if (!Block.matchingIdAndGenStamp(previousBlock, lastBlockInFile)) {
// The block that the client claims is the current last block
@ -2294,7 +2300,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
}
//check lease
final INodeFileUnderConstruction file = checkLease(src, clientName);
final INodeFileUnderConstruction file = checkLease(
src, clientName, dir.getINode(src));
clientnode = file.getClientNode();
preferredblocksize = file.getPreferredBlockSize();
@ -2340,7 +2347,9 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
throw new SafeModeException("Cannot abandon block " + b +
" for fle" + src, safeMode);
}
INodeFileUnderConstruction file = checkLease(src, holder);
final INodesInPath inodesInPath = checkLease(src, holder);
final INodeFileUnderConstruction file
= (INodeFileUnderConstruction)inodesInPath.getINode(0);
dir.removeBlock(src, file, ExtendedBlock.getLocalBlock(b));
if(NameNode.stateChangeLog.isDebugEnabled()) {
NameNode.stateChangeLog.debug("BLOCK* NameSystem.abandonBlock: "
@ -2357,11 +2366,13 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
return true;
}
// make sure that we still have the lease on this file.
private INodeFileUnderConstruction checkLease(String src, String holder)
/** make sure that we still have the lease on this file. */
private INodesInPath checkLease(String src, String holder)
throws LeaseExpiredException, UnresolvedLinkException {
assert hasReadOrWriteLock();
return checkLease(src, holder, dir.getINode(src));
final INodesInPath inodesInPath = dir.getINodesInPath(src);
checkLease(src, holder, inodesInPath.getINode(0));
return inodesInPath;
}
private INodeFileUnderConstruction checkLease(String src, String holder,
@ -2424,9 +2435,11 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
throw new SafeModeException("Cannot complete file " + src, safeMode);
}
INodeFileUnderConstruction pendingFile;
final INodesInPath inodesInPath;
final INodeFileUnderConstruction pendingFile;
try {
pendingFile = checkLease(src, holder);
inodesInPath = checkLease(src, holder);
pendingFile = (INodeFileUnderConstruction)inodesInPath.getINode(0);
} catch (LeaseExpiredException lee) {
final INode inode = dir.getINode(src);
if (inode != null && inode instanceof INodeFile && !inode.isUnderConstruction()) {
@ -2454,7 +2467,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
return false;
}
finalizeINodeFileUnderConstruction(src, pendingFile);
finalizeINodeFileUnderConstruction(src, pendingFile,
inodesInPath.getLatestSnapshot());
NameNode.stateChangeLog.info("DIR* completeFile: " + src + " is closed by "
+ holder);
@ -2595,15 +2609,15 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
if (isPermissionEnabled) {
//We should not be doing this. This is move() not renameTo().
//but for now,
//NOTE: yes, this is bad! it's assuming much lower level behavior
// of rewriting the dst
String actualdst = dir.isDir(dst)?
dst + Path.SEPARATOR + new Path(src).getName(): dst;
checkParentAccess(src, FsAction.WRITE);
checkAncestorAccess(actualdst, FsAction.WRITE);
}
HdfsFileStatus dinfo = dir.getFileInfo(dst, false);
if (dir.renameTo(src, dst)) {
unprotectedChangeLease(src, dst, dinfo); // update lease with new filename
return true;
}
return false;
@ -2654,9 +2668,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
checkAncestorAccess(dst, FsAction.WRITE);
}
HdfsFileStatus dinfo = dir.getFileInfo(dst, false);
dir.renameTo(src, dst, options);
unprotectedChangeLease(src, dst, dinfo); // update lease with new filename
}
/**
@ -3036,7 +3048,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
if (isInSafeMode()) {
throw new SafeModeException("Cannot fsync file " + src, safeMode);
}
INodeFileUnderConstruction pendingFile = checkLease(src, clientName);
final INodeFileUnderConstruction pendingFile = checkLease(
src, clientName, dir.getINode(src));
if (lastBlockLength > 0) {
pendingFile.updateLengthOfLastBlock(lastBlockLength);
}
@ -3068,8 +3081,9 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
assert !isInSafeMode();
assert hasWriteLock();
final INodesInPath inodesInPath = dir.getINodesInPath(src);
final INodeFileUnderConstruction pendingFile
= INodeFileUnderConstruction.valueOf(dir.getINode(src), src);
= INodeFileUnderConstruction.valueOf(inodesInPath.getINode(0), src);
int nrBlocks = pendingFile.numBlocks();
BlockInfo[] blocks = pendingFile.getBlocks();
@ -3086,7 +3100,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// If there are no incomplete blocks associated with this file,
// then reap lease immediately and close the file.
if(nrCompleteBlocks == nrBlocks) {
finalizeINodeFileUnderConstruction(src, pendingFile);
finalizeINodeFileUnderConstruction(src, pendingFile,
inodesInPath.getLatestSnapshot());
NameNode.stateChangeLog.warn("BLOCK*"
+ " internalReleaseLease: All existing blocks are COMPLETE,"
+ " lease removed, file closed.");
@ -3134,7 +3149,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// Close file if committed blocks are minimally replicated
if(penultimateBlockMinReplication &&
blockManager.checkMinReplication(lastBlock)) {
finalizeINodeFileUnderConstruction(src, pendingFile);
finalizeINodeFileUnderConstruction(src, pendingFile,
inodesInPath.getLatestSnapshot());
NameNode.stateChangeLog.warn("BLOCK*"
+ " internalReleaseLease: Committed blocks are minimally replicated,"
+ " lease removed, file closed.");
@ -3212,7 +3228,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
}
private void finalizeINodeFileUnderConstruction(String src,
INodeFileUnderConstruction pendingFile)
INodeFileUnderConstruction pendingFile, Snapshot latestSnapshot)
throws IOException, UnresolvedLinkException {
assert hasWriteLock();
leaseManager.removeLease(pendingFile.getClientName(), src);
@ -3220,7 +3236,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// The file is no longer pending.
// Create permanent INode, update blocks
INodeFile newFile = pendingFile.convertToInodeFile();
dir.replaceNode(src, pendingFile, newFile);
dir.replaceNode(src, pendingFile, newFile, latestSnapshot);
// close file and persist block allocations for this file
dir.closeFile(src, newFile);
@ -3312,7 +3328,8 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
commitOrCompleteLastBlock(pendingFile, storedBlock);
//remove lease, close file
finalizeINodeFileUnderConstruction(src, pendingFile);
finalizeINodeFileUnderConstruction(src, pendingFile,
INodeDirectorySnapshottable.findLatestSnapshot(pendingFile));
} else {
// If this commit does not want to close the file, persist blocks
dir.persistBlocks(src, pendingFile);
@ -3483,9 +3500,9 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
private NNHAStatusHeartbeat createHaStatusHeartbeat() {
HAState state = haContext.getState();
NNHAStatusHeartbeat.State hbState;
if (state instanceof ActiveState) {
if (state.getServiceState() == HAServiceState.ACTIVE) {
hbState = NNHAStatusHeartbeat.State.ACTIVE;
} else if (state instanceof StandbyState) {
} else if (state.getServiceState() == HAServiceState.STANDBY) {
hbState = NNHAStatusHeartbeat.State.STANDBY;
} else {
throw new AssertionError("Invalid state: " + state.getClass());
@ -4939,31 +4956,9 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// rename was successful. If any part of the renamed subtree had
// files that were being written to, update with new filename.
void unprotectedChangeLease(String src, String dst, HdfsFileStatus dinfo) {
String overwrite;
String replaceBy;
void unprotectedChangeLease(String src, String dst) {
assert hasWriteLock();
boolean destinationExisted = true;
if (dinfo == null) {
destinationExisted = false;
}
if (destinationExisted && dinfo.isDir()) {
Path spath = new Path(src);
Path parent = spath.getParent();
if (parent.isRoot()) {
overwrite = parent.toString();
} else {
overwrite = parent.toString() + Path.SEPARATOR;
}
replaceBy = dst + Path.SEPARATOR;
} else {
overwrite = src;
replaceBy = dst;
}
leaseManager.changeLease(src, dst, overwrite, replaceBy);
leaseManager.changeLease(src, dst);
}
/**
@ -4974,19 +4969,13 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
// lock on our behalf. If we took the read lock here, we could block
// for fairness if a writer is waiting on the lock.
synchronized (leaseManager) {
out.writeInt(leaseManager.countPath()); // write the size
for (Lease lease : leaseManager.getSortedLeases()) {
for(String path : lease.getPaths()) {
// verify that path exists in namespace
final INodeFileUnderConstruction cons;
try {
cons = INodeFileUnderConstruction.valueOf(dir.getINode(path), path);
} catch (UnresolvedLinkException e) {
throw new AssertionError("Lease files should reside on this FS");
}
FSImageSerialization.writeINodeUnderConstruction(out, cons, path);
}
Map<String, INodeFileUnderConstruction> nodes =
leaseManager.getINodesUnderConstruction();
out.writeInt(nodes.size()); // write the size
for (Map.Entry<String, INodeFileUnderConstruction> entry
: nodes.entrySet()) {
FSImageSerialization.writeINodeUnderConstruction(
out, entry.getValue(), entry.getKey());
}
}
}

59
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/INode.java
View File

@ -33,6 +33,7 @@ import org.apache.hadoop.hdfs.DFSUtil;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockCollection;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.namenode.snapshot.Snapshot;
import org.apache.hadoop.hdfs.util.ReadOnlyList;
import org.apache.hadoop.util.StringUtils;
@ -46,8 +47,33 @@ import com.google.common.primitives.SignedBytes;
*/
@InterfaceAudience.Private
public abstract class INode implements Comparable<byte[]> {
/** A dummy INode which can be used as a probe object. */
public static final INode DUMMY = new INode() {
@Override
int collectSubtreeBlocksAndClear(BlocksMapUpdateInfo info) {
throw new UnsupportedOperationException();
}
@Override
long[] computeContentSummary(long[] summary) {
throw new UnsupportedOperationException();
}
@Override
DirCounts spaceConsumedInTree(DirCounts counts) {
throw new UnsupportedOperationException();
}
};
static final ReadOnlyList<INode> EMPTY_READ_ONLY_LIST
= ReadOnlyList.Util.emptyList();
/**
* Assert that the snapshot parameter must be null since
* this class only take care current state.
* Subclasses should override the methods for handling the snapshot states.
*/
static void assertNull(Snapshot snapshot) {
if (snapshot != null) {
throw new AssertionError("snapshot is not null: " + snapshot);
}
}
/** Wrapper of two counters for namespace consumed and diskspace consumed. */
@ -120,9 +146,12 @@ public abstract class INode implements Comparable<byte[]> {
* should not modify it.
*/
private long permission = 0L;
protected INodeDirectory parent = null;
protected long modificationTime = 0L;
protected long accessTime = 0L;
INodeDirectory parent = null;
private long modificationTime = 0L;
private long accessTime = 0L;
/** For creating the a {@link #DUMMY} object. */
private INode() {}
private INode(byte[] name, long permission, INodeDirectory parent,
long modificationTime, long accessTime) {
@ -149,8 +178,8 @@ public abstract class INode implements Comparable<byte[]> {
/** @param other Other node to be copied */
INode(INode other) {
this(other.getLocalNameBytes(), other.permission, other.getParent(),
other.getModificationTime(), other.getAccessTime());
this(other.name, other.permission, other.parent,
other.modificationTime, other.accessTime);
}
/**
@ -290,13 +319,13 @@ public abstract class INode implements Comparable<byte[]> {
* Set local file name
*/
public void setLocalName(String name) {
this.name = DFSUtil.string2Bytes(name);
setLocalName(DFSUtil.string2Bytes(name));
}
/**
* Set local file name
*/
void setLocalName(byte[] name) {
public void setLocalName(byte[] name) {
this.name = name;
}
@ -316,7 +345,7 @@ public abstract class INode implements Comparable<byte[]> {
* Get parent directory
* @return parent INode
*/
INodeDirectory getParent() {
public INodeDirectory getParent() {
return this.parent;
}
@ -336,17 +365,21 @@ public abstract class INode implements Comparable<byte[]> {
/**
* Set last modification time of inode.
*/
public void setModificationTime(long modtime) {
public void updateModificationTime(long modtime) {
assert isDirectory();
if (this.modificationTime <= modtime) {
this.modificationTime = modtime;
setModificationTime(modtime);
}
}
void cloneModificationTime(INode that) {
this.modificationTime = that.modificationTime;
}
/**
* Always set the last modification time of inode.
*/
void setModificationTimeForce(long modtime) {
void setModificationTime(long modtime) {
this.modificationTime = modtime;
}
@ -431,11 +464,11 @@ public abstract class INode implements Comparable<byte[]> {
return buf.toString();
}
public boolean removeNode() {
public boolean removeNode(Snapshot latestSnapshot) {
if (parent == null) {
return false;
} else {
parent.removeChild(this);
parent.removeChild(this, latestSnapshot);
parent = null;
return true;
}

90
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/INodeDirectory.java
View File

@ -99,32 +99,56 @@ public class INodeDirectory extends INode {
}
}
private int searchChildren(INode inode) {
public int searchChildren(INode inode) {
return Collections.binarySearch(children, inode.getLocalNameBytes());
}
INode removeChild(INode node) {
public int searchChildrenForExistingINode(INode inode) {
final int i = searchChildren(inode);
if (i < 0) {
throw new AssertionError("Child not found: inode=" + inode);
}
return i;
}
public INode removeChild(INode node, Snapshot latestSnapshot) {
assertChildrenNonNull();
if (latestSnapshot != null) {
final INodeDirectoryWithSnapshot dir
= INodeDirectoryWithSnapshot.replaceDir(this, latestSnapshot);
return dir.removeChild(node, latestSnapshot);
}
final int i = searchChildren(node);
return i >= 0? children.remove(i): null;
}
/** Replace a child that has the same name as newChild by newChild.
*
* @param newChild Child node to be added
*/
void replaceChild(INode newChild) {
public INode replaceChild(INodeDirectory newChild, Snapshot latestSnapshot) {
assertChildrenNonNull();
final int low = searchChildren(newChild);
if (low>=0) { // an old child exists so replace by the newChild
children.set(low, newChild);
} else {
throw new IllegalArgumentException("No child exists to be replaced");
}
if (latestSnapshot != null) {
final INodeDirectoryWithSnapshot dir
= INodeDirectoryWithSnapshot.replaceDir(this, latestSnapshot);
return dir.replaceChild(newChild, latestSnapshot);
}
private INode getChild(byte[] name, Snapshot snapshot) {
// find the old child and replace it
final int low = searchChildrenForExistingINode(newChild);
final INode oldChild = children.set(low, newChild);
// set the parent of the children of the child.
for(INode i : newChild.getChildrenList(null)) {
i.parent = newChild;
}
return oldChild;
}
public INode getChild(byte[] name, Snapshot snapshot) {
assertNull(snapshot);
final ReadOnlyList<INode> c = getChildrenList(snapshot);
final int i = ReadOnlyList.Util.binarySearch(c, name);
return i < 0? null: c.get(i);
@ -361,7 +385,14 @@ public class INodeDirectory extends INode {
* @return false if the child with this name already exists;
* otherwise, return true;
*/
public boolean addChild(final INode node, final boolean setModTime) {
public boolean addChild(final INode node, final boolean setModTime,
Snapshot latestSnapshot) {
if (latestSnapshot != null) {
final INodeDirectoryWithSnapshot dir
= INodeDirectoryWithSnapshot.replaceDir(this, latestSnapshot);
return dir.addChild(node, setModTime, latestSnapshot);
}
if (children == null) {
children = new ArrayList<INode>(DEFAULT_FILES_PER_DIRECTORY);
}
@ -372,8 +403,9 @@ public class INodeDirectory extends INode {
node.parent = this;
children.add(-low - 1, node);
// update modification time of the parent directory
if (setModTime)
setModificationTime(node.getModificationTime());
if (setModTime) {
updateModificationTime(node.getModificationTime());
}
if (node.getGroupName() == null) {
node.setGroup(getGroupName());
}
@ -400,18 +432,26 @@ public class INodeDirectory extends INode {
}
newNode.setLocalName(pathComponents[pathComponents.length - 1]);
// insert into the parent children list
INodeDirectory parent = getParent(pathComponents);
return parent.addChild(newNode, true);
INodesInPath inodes = getExistingPathINodes(pathComponents, 2, false);
INodeDirectory parent = INodeDirectory.valueOf(inodes.inodes[0], pathComponents);
return parent.addChild(newNode, true, inodes.getLatestSnapshot());
}
INodeDirectory getParent(byte[][] pathComponents
) throws FileNotFoundException, PathIsNotDirectoryException,
UnresolvedLinkException {
return (INodeDirectory)getParentINodesInPath(pathComponents).getINode(0);
}
INodesInPath getParentINodesInPath(byte[][] pathComponents
) throws FileNotFoundException, PathIsNotDirectoryException,
UnresolvedLinkException {
if (pathComponents.length < 2) // add root
return null;
// Gets the parent INode
INodesInPath inodes = getExistingPathINodes(pathComponents, 2, false);
return INodeDirectory.valueOf(inodes.inodes[0], pathComponents);
INodeDirectory.valueOf(inodes.inodes[0], pathComponents);
return inodes;
}
@Override
@ -462,10 +502,11 @@ public class INodeDirectory extends INode {
* Note that the returned list is never null.
*/
public ReadOnlyList<INode> getChildrenList(final Snapshot snapshot) {
//TODO: use snapshot to select children list
assertNull(snapshot);
return children == null ? EMPTY_READ_ONLY_LIST
: ReadOnlyList.Util.asReadOnlyList(children);
}
/** Set the children list. */
public void setChildren(List<INode> children) {
this.children = children;
@ -490,7 +531,7 @@ public class INodeDirectory extends INode {
* {@link INodeDirectory#getExistingPathINodes(byte[][], int, boolean)}.
* Contains INodes information resolved from a given path.
*/
static class INodesInPath {
public static class INodesInPath {
/**
* Array with the specified number of INodes resolved for a given path.
*/
@ -570,7 +611,7 @@ public class INodeDirectory extends INode {
}
/** @return the i-th inode. */
INode getINode(int i) {
public INode getINode(int i) {
return inodes[i];
}
@ -676,4 +717,13 @@ public class INodeDirectory extends INode {
}
}
}
/**
* Get last modification time of inode.
* @return access time
*/
public long getModificationTime(Snapshot snapshot) {
assertNull(snapshot);
return getModificationTime();
}
}

16
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/INodeFile.java
View File

@ -89,16 +89,22 @@ public class INodeFile extends INode implements BlockCollection {
INodeFile(PermissionStatus permissions, BlockInfo[] blklist,
short replication, long modificationTime,
long atime, long preferredBlockSize) {
super(permissions, modificationTime, atime);
this(null, permissions, modificationTime, atime, blklist, replication,
preferredBlockSize);
}
INodeFile(byte[] name, PermissionStatus permissions, long mtime, long atime,
BlockInfo[] blklist, short replication, long preferredBlockSize) {
super(name, permissions, null, mtime, atime);
header = HeaderFormat.combineReplication(header, replication);
header = HeaderFormat.combinePreferredBlockSize(header, preferredBlockSize);
this.blocks = blklist;
}
protected INodeFile(INodeFile f) {
this(f.getPermissionStatus(), f.getBlocks(), f.getFileReplication(),
f.getModificationTime(), f.getAccessTime(), f.getPreferredBlockSize());
this.setLocalName(f.getLocalNameBytes());
protected INodeFile(INodeFile that) {
super(that);
this.header = that.header;
this.blocks = that.blocks;
}
/** @return true unconditionally. */

5
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/INodeFileUnderConstruction.java
View File

@ -72,9 +72,8 @@ public class INodeFileUnderConstruction extends INodeFile implements MutableBloc
String clientName,
String clientMachine,
DatanodeDescriptor clientNode) {
super(perm, blocks, blockReplication, modificationTime, modificationTime,
preferredBlockSize);
setLocalName(name);
super(name, perm, modificationTime, modificationTime,
blocks, blockReplication, preferredBlockSize);
this.clientName = clientName;
this.clientMachine = clientMachine;
this.clientNode = clientNode;

33
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/LeaseManager.java
View File

@ -331,22 +331,19 @@ public class LeaseManager {
}
}
synchronized void changeLease(String src, String dst,
String overwrite, String replaceBy) {
synchronized void changeLease(String src, String dst) {
if (LOG.isDebugEnabled()) {
LOG.debug(getClass().getSimpleName() + ".changelease: " +
" src=" + src + ", dest=" + dst +
", overwrite=" + overwrite +
", replaceBy=" + replaceBy);
" src=" + src + ", dest=" + dst);
}
final int len = overwrite.length();
final int len = src.length();
for(Map.Entry<String, Lease> entry
: findLeaseWithPrefixPath(src, sortedLeasesByPath).entrySet()) {
final String oldpath = entry.getKey();
final Lease lease = entry.getValue();
//overwrite must be a prefix of oldpath
final String newpath = replaceBy + oldpath.substring(len);
// replace stem of src with new destination
final String newpath = dst + oldpath.substring(len);
if (LOG.isDebugEnabled()) {
LOG.debug("changeLease: replacing " + oldpath + " with " + newpath);
}
@ -429,6 +426,26 @@ public class LeaseManager {
}
}
/**
* Get the list of inodes corresponding to valid leases.
* @return list of inodes
* @throws UnresolvedLinkException
*/
Map<String, INodeFileUnderConstruction> getINodesUnderConstruction() {
Map<String, INodeFileUnderConstruction> inodes =
new TreeMap<String, INodeFileUnderConstruction>();
for (String p : sortedLeasesByPath.keySet()) {
// verify that path exists in namespace
try {
INode node = fsnamesystem.dir.getINode(p);
inodes.put(p, INodeFileUnderConstruction.valueOf(node, p));
} catch (IOException ioe) {
LOG.error(ioe);
}
}
return inodes;
}
/** Check the leases beginning from the oldest.
* @return true is sync is needed.
*/

12
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NameNode.java
View File

@ -598,11 +598,7 @@ public class NameNode {
String nsId = getNameServiceId(conf);
String namenodeId = HAUtil.getNameNodeId(conf, nsId);
this.haEnabled = HAUtil.isHAEnabled(conf, nsId);
if (!haEnabled) {
state = ACTIVE_STATE;
} else {
state = STANDBY_STATE;
}
state = createHAState();
this.allowStaleStandbyReads = HAUtil.shouldAllowStandbyReads(conf);
this.haContext = createHAContext();
try {
@ -619,6 +615,10 @@ public class NameNode {
}
}
protected HAState createHAState() {
return !haEnabled ? ACTIVE_STATE : STANDBY_STATE;
}
protected HAContext createHAContext() {
return new NameNodeHAContext();
}
@ -1298,7 +1298,7 @@ public class NameNode {
* before exit.
* @throws ExitException thrown only for testing.
*/
private synchronized void doImmediateShutdown(Throwable t)
protected synchronized void doImmediateShutdown(Throwable t)
throws ExitException {
String message = "Error encountered requiring NN shutdown. " +
"Shutting down immediately.";

2
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NamenodeJspHelper.java
View File

@ -102,7 +102,7 @@ class NamenodeJspHelper {
long usedNonHeap = (totalNonHeap * 100) / commitedNonHeap;
String str = "<div>" + inodes + " files and directories, " + blocks + " blocks = "
+ (inodes + blocks) + " total";
+ (inodes + blocks) + " total filesystem objects";
if (maxobjects != 0) {
long pct = ((inodes + blocks) * 100) / maxobjects;
str += " / " + maxobjects + " (" + pct + "%)";

17
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/snapshot/INodeDirectorySnapshottable.java
View File

@ -66,6 +66,19 @@ public class INodeDirectorySnapshottable extends INodeDirectoryWithQuota {
return (INodeDirectorySnapshottable)dir;
}
public static Snapshot findLatestSnapshot(INode inode) {
Snapshot latest = null;
for(; inode != null; inode = inode.getParent()) {
if (inode instanceof INodeDirectorySnapshottable) {
final Snapshot s = ((INodeDirectorySnapshottable)inode).getLastSnapshot();
if (Snapshot.ID_COMPARATOR.compare(latest, s) < 0) {
latest = s;
}
}
}
return latest;
}
/** Snapshots of this directory in ascending order of snapshot id. */
private final List<Snapshot> snapshots = new ArrayList<Snapshot>();
/** Snapshots of this directory in ascending order of snapshot names. */
@ -196,8 +209,8 @@ public class INodeDirectorySnapshottable extends INodeDirectoryWithQuota {
//set modification time
final long timestamp = Time.now();
s.getRoot().setModificationTime(timestamp);
setModificationTime(timestamp);
s.getRoot().updateModificationTime(timestamp);
updateModificationTime(timestamp);
return s;
}

225
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/snapshot/INodeDirectoryWithSnapshot.java
View File

@ -19,10 +19,15 @@ package org.apache.hadoop.hdfs.server.namenode.snapshot;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;
import org.apache.hadoop.HadoopIllegalArgumentException;
import org.apache.hadoop.hdfs.server.namenode.INode;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory;
import org.apache.hadoop.hdfs.util.ReadOnlyList;
import com.google.common.base.Preconditions;
/** The directory with snapshots. */
public class INodeDirectoryWithSnapshot extends INodeDirectory {
@ -251,8 +256,222 @@ public class INodeDirectoryWithSnapshot extends INodeDirectory {
}
}
INodeDirectoryWithSnapshot(String name, INodeDirectory dir) {
super(name, dir.getPermissionStatus());
parent = dir;
private class SnapshotDiff implements Comparable<Snapshot> {
/** The snapshot will obtain after applied this diff. */
final Snapshot snapshot;
/** The size of the children list which is never changed. */
final int size;
/**
* Posterior diff is the diff happened after this diff.
* The posterior diff should be first applied to obtain the posterior
* snapshot and then apply this diff in order to obtain this snapshot.
* If the posterior diff is null, the posterior state is the current state.
*/
private SnapshotDiff posteriorDiff;
/** The data of this diff. */
private final Diff diff = new Diff();
/** The snapshot version of the inode. */
private INode snapshotINode;
SnapshotDiff(Snapshot snapshot, int size) {
if (size < 0) {
throw new HadoopIllegalArgumentException("size = " + size + " < 0");
}
this.snapshot = snapshot;
this.size = size;
}
@Override
public int compareTo(final Snapshot that_snapshot) {
return Snapshot.ID_COMPARATOR.compare(this.snapshot, that_snapshot);
}
/**
* @return The children list of a directory in a snapshot.
* Since the snapshot is read-only, the logical view of the list is
* never changed although the internal data structure may mutate.
*/
ReadOnlyList<INode> getChildrenList() {
return new ReadOnlyList<INode>() {
private List<INode> children = null;
private List<INode> initChildren() {
if (children == null) {
final ReadOnlyList<INode> posterior = posteriorDiff != null?
posteriorDiff.getChildrenList()
: INodeDirectoryWithSnapshot.this.getChildrenList(null);
children = diff.apply2Current(ReadOnlyList.Util.asList(posterior));
}
return children;
}
@Override
public Iterator<INode> iterator() {
return initChildren().iterator();
}
@Override
public boolean isEmpty() {
return size == 0;
}
@Override
public int size() {
return size;
}
@Override
public INode get(int i) {
return initChildren().get(i);
}
};
}
INode getChild(byte[] name) {
final INode i = diff.accessPrevious(name, INode.DUMMY);
if (i != INode.DUMMY) {
// this diff is able to find it
return i;
} else {
// should return the posterior INode.
return posteriorDiff != null? posteriorDiff.getChild(name)
: INodeDirectoryWithSnapshot.this.getChild(name, null);
}
}
}
/** Replace the given directory to an {@link INodeDirectoryWithSnapshot}. */
public static INodeDirectoryWithSnapshot replaceDir(INodeDirectory oldDir,
Snapshot latestSnapshot) {
Preconditions.checkArgument(!(oldDir instanceof INodeDirectoryWithSnapshot),
"oldDir is already an INodeDirectoryWithSnapshot, oldDir=%s", oldDir);
final INodeDirectory parent = oldDir.getParent();
Preconditions.checkArgument(parent != null,
"parent is null, oldDir=%s", oldDir);
final INodeDirectoryWithSnapshot newDir = new INodeDirectoryWithSnapshot(
oldDir, latestSnapshot);
parent.replaceChild(newDir, null);
return newDir;
}
/** Diff list sorted by snapshot IDs, i.e. in chronological order. */
private final List<SnapshotDiff> diffs = new ArrayList<SnapshotDiff>();
INodeDirectoryWithSnapshot(INodeDirectory that, Snapshot s) {
super(that);
// add a diff for the snapshot
addSnapshotDiff(s, that.getChildrenList(null).size());
}
INodeDirectoryWithSnapshot(String name, INodeDirectory dir, Snapshot s) {
this(dir, s);
setLocalName(name);
setParent(dir);
}
SnapshotDiff addSnapshotDiff(Snapshot snapshot, int childrenSize) {
final SnapshotDiff d = new SnapshotDiff(snapshot, childrenSize);
diffs.add(d);
return d;
}
/**
* Check if the latest snapshot diff exist. If not, add it.
* @return the latest snapshot diff, which is never null.
*/
private SnapshotDiff checkAndAddLatestSnapshotDiff(Snapshot latest) {
final SnapshotDiff last = getLastSnapshotDiff();
if (last != null && last.snapshot.equals(latest)) {
return last;
}
final int size = getChildrenList(null).size();
final SnapshotDiff d = addSnapshotDiff(latest, size);
if (last != null) {
last.posteriorDiff = d;
}
return d;
}
Diff getLatestDiff(Snapshot latest) {
return checkAndAddLatestSnapshotDiff(latest).diff;
}
/**
* @return the diff corresponding to the snapshot.
* When the diff is not found, it means that the current state and
* the snapshot state are the same.
*/
SnapshotDiff getSnapshotDiff(Snapshot snapshot) {
if (snapshot == null) {
return null;
}
final int i = Collections.binarySearch(diffs, snapshot);
if (i >= 0) {
// exact match
return diffs.get(i);
} else {
// Exact match not found means that there were no changes between
// given snapshot and the next state so that the diff for the given
// snapshot is not recorded. Thus, use the next state.
final int j = -i - 1;
return j < diffs.size()? diffs.get(j): null;
}
}
SnapshotDiff getLastSnapshotDiff() {
return diffs.get(diffs.size() - 1);
}
@Override
public ReadOnlyList<INode> getChildrenList(Snapshot snapshot) {
final SnapshotDiff diff = getSnapshotDiff(snapshot);
if (diff != null) {
return diff.getChildrenList();
}
return super.getChildrenList(null);
}
@Override
public INode getChild(byte[] name, Snapshot snapshot) {
final SnapshotDiff diff = getSnapshotDiff(snapshot);
if (diff != null) {
return diff.getChild(name);
}
return super.getChild(name, null);
}
@Override
public boolean addChild(INode inode, boolean setModTime,
Snapshot latestSnapshot) {
getLatestDiff(latestSnapshot).create(inode);
return super.addChild(inode, setModTime, null);
}
@Override
public INode removeChild(INode inode, Snapshot latestSnapshot) {
getLatestDiff(latestSnapshot).delete(inode);
return super.removeChild(inode, null);
}
@Override
public INode replaceChild(INodeDirectory newChild, Snapshot latestSnapshot) {
final INode oldChild = super.replaceChild(newChild, null);
final Diff diff = getLatestDiff(latestSnapshot);
diff.delete(oldChild);
diff.create(newChild);
return oldChild;
}
@Override
public long getModificationTime(Snapshot snapshot) {
final SnapshotDiff diff = getSnapshotDiff(snapshot);
if (diff != null) {
return diff.snapshotINode.getModificationTime();
}
return getModificationTime();
}
}

2
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/snapshot/INodeFileWithLink.java
View File

@ -109,7 +109,7 @@ public class INodeFileWithLink extends INodeFile {
this.setFileReplication(maxReplication);
this.next = null;
// clear parent
parent = null;
setParent(null);
}
return 1;
}

17
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/snapshot/Snapshot.java
View File

@ -44,7 +44,7 @@ public class Snapshot implements Comparable<byte[]> {
Snapshot(int id, String name, INodeDirectorySnapshottable dir) {
this.id = id;
this.root = new INodeDirectoryWithSnapshot(name, dir);
this.root = new INodeDirectoryWithSnapshot(name, dir, this);
}
/** @return the root directory of the snapshot. */
@ -57,6 +57,21 @@ public class Snapshot implements Comparable<byte[]> {
return root.compareTo(bytes);
}
@Override
public boolean equals(Object that) {
if (this == that) {
return true;
} else if (that == null || !(that instanceof Snapshot)) {
return false;
}
return this.id == ((Snapshot)that).id;
}
@Override
public int hashCode() {
return id;
}
@Override
public String toString() {
return getClass().getSimpleName() + ":" + root.getLocalName();

99
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/snapshot/SnapshotManager.java
View File

@ -24,12 +24,8 @@ import java.util.concurrent.atomic.AtomicInteger;
import org.apache.hadoop.hdfs.server.namenode.FSDirectory;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.INode;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory;
import org.apache.hadoop.hdfs.server.namenode.INodeFile;
import org.apache.hadoop.hdfs.server.namenode.INodeFileUnderConstruction;
import org.apache.hadoop.hdfs.server.namenode.INodeSymlink;
import org.apache.hadoop.hdfs.util.ReadOnlyList;
import org.apache.hadoop.hdfs.server.namenode.INodeDirectory.INodesInPath;
/**
* Manage snapshottable directories and their snapshots.
@ -68,7 +64,9 @@ public class SnapshotManager implements SnapshotStats {
*/
public void setSnapshottable(final String path, final int snapshotQuota
) throws IOException {
final INodeDirectory d = INodeDirectory.valueOf(fsdir.getINode(path), path);
final INodesInPath inodesInPath = fsdir.getINodesInPath(path);
final INodeDirectory d = INodeDirectory.valueOf(
inodesInPath.getINode(0), path);
if (d.isSnapshottable()) {
//The directory is already a snapshottable directory.
((INodeDirectorySnapshottable)d).setSnapshotQuota(snapshotQuota);
@ -77,7 +75,7 @@ public class SnapshotManager implements SnapshotStats {
final INodeDirectorySnapshottable s
= INodeDirectorySnapshottable.newInstance(d, snapshotQuota);
fsdir.replaceINodeDirectory(path, d, s);
fsdir.replaceINodeDirectory(path, d, s, inodesInPath.getLatestSnapshot());
snapshottables.add(s);
numSnapshottableDirs.getAndIncrement();
@ -90,15 +88,16 @@ public class SnapshotManager implements SnapshotStats {
*/
public void resetSnapshottable(final String path
) throws IOException {
final INodesInPath inodesInPath = fsdir.getINodesInPath(path);
final INodeDirectorySnapshottable s = INodeDirectorySnapshottable.valueOf(
fsdir.getINode(path), path);
inodesInPath.getINode(0), path);
if (s.getNumSnapshots() > 0) {
throw new SnapshotException("The directory " + path + " has snapshot(s). "
+ "Please redo the operation after removing all the snapshots.");
}
final INodeDirectory d = new INodeDirectory(s);
fsdir.replaceINodeDirectory(path, s, d);
fsdir.replaceINodeDirectory(path, s, d, inodesInPath.getLatestSnapshot());
snapshottables.remove(s);
numSnapshottableDirs.getAndDecrement();
@ -121,8 +120,7 @@ public class SnapshotManager implements SnapshotStats {
// Find the source root directory path where the snapshot is taken.
final INodeDirectorySnapshottable srcRoot
= INodeDirectorySnapshottable.valueOf(fsdir.getINode(path), path);
final Snapshot s = srcRoot.addSnapshot(snapshotID, snapshotName);
new SnapshotCreation().processRecursively(srcRoot, s.getRoot());
srcRoot.addSnapshot(snapshotID, snapshotName);
//create success, update id
snapshotID++;
@ -154,85 +152,6 @@ public class SnapshotManager implements SnapshotStats {
srcRoot.renameSnapshot(path, oldSnapshotName, newSnapshotName);
}
/**
* Create a snapshot of subtrees by recursively coping the directory
* structure from the source directory to the snapshot destination directory.
* This creation algorithm requires O(N) running time and O(N) memory,
* where N = # files + # directories + # symlinks.
*/
class SnapshotCreation {
/** Process snapshot creation recursively. */
private void processRecursively(final INodeDirectory srcDir,
final INodeDirectory dstDir) throws IOException {
final ReadOnlyList<INode> children = srcDir.getChildrenList(null);
if (!children.isEmpty()) {
final List<INode> inodes = new ArrayList<INode>(children.size());
for(final INode c : new ArrayList<INode>(ReadOnlyList.Util.asList(children))) {
final INode i;
if (c == null) {
i = null;
} else if (c instanceof INodeDirectory) {
//also handle INodeDirectoryWithQuota
i = processINodeDirectory((INodeDirectory)c);
} else if (c instanceof INodeFileUnderConstruction) {
//TODO: support INodeFileUnderConstruction
throw new IOException("Not yet supported.");
} else if (c instanceof INodeFile) {
i = processINodeFile(srcDir, (INodeFile)c);
} else if (c instanceof INodeSymlink) {
i = new INodeSymlink((INodeSymlink)c);
} else {
throw new AssertionError("Unknow INode type: " + c.getClass()
+ ", inode = " + c);
}
i.setParent(dstDir);
inodes.add(i);
}
dstDir.setChildren(inodes);
}
}
/**
* Create destination INodeDirectory and make the recursive call.
* @return destination INodeDirectory.
*/
private INodeDirectory processINodeDirectory(final INodeDirectory srcChild
) throws IOException {
final INodeDirectory dstChild = new INodeDirectory(srcChild);
dstChild.setChildren(null);
processRecursively(srcChild, dstChild);
return dstChild;
}
/**
* Create destination INodeFileSnapshot and update source INode type.
* @return destination INodeFileSnapshot.
*/
private INodeFileSnapshot processINodeFile(final INodeDirectory parent,
final INodeFile file) {
final INodeFileSnapshot snapshot = new INodeFileSnapshot(
file, file.computeFileSize(true));
final INodeFileWithLink srcWithLink;
//check source INode type
if (file instanceof INodeFileWithLink) {
srcWithLink = (INodeFileWithLink)file;
} else {
//source is an INodeFile, replace the source.
srcWithLink = new INodeFileWithLink(file);
file.removeNode();
parent.addChild(srcWithLink, false);
//update block map
namesystem.getBlockManager().addBlockCollection(srcWithLink);
}
//insert the snapshot to src's linked list.
srcWithLink.insert(snapshot);
return snapshot;
}
}
@Override
public long getNumSnapshottableDirs() {
return numSnapshottableDirs.get();

96
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestLease.java
View File

@ -30,7 +30,9 @@ import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Options;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hdfs.protocol.ClientProtocol;
import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
@ -49,6 +51,10 @@ public class TestLease {
).getLeaseByPath(src.toString()) != null;
}
static int leaseCount(MiniDFSCluster cluster) {
return NameNodeAdapter.getLeaseManager(cluster.getNamesystem()).countLease();
}
static final String dirString = "/test/lease";
final Path dir = new Path(dirString);
static final Log LOG = LogFactory.getLog(TestLease.class);
@ -126,6 +132,96 @@ public class TestLease {
}
}
@Test
public void testLeaseAfterRename() throws Exception {
MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf).numDataNodes(2).build();
try {
Path p = new Path("/test-file");
Path d = new Path("/test-d");
Path d2 = new Path("/test-d-other");
// open a file to get a lease
FileSystem fs = cluster.getFileSystem();
FSDataOutputStream out = fs.create(p);
out.writeBytes("something");
//out.hsync();
Assert.assertTrue(hasLease(cluster, p));
Assert.assertEquals(1, leaseCount(cluster));
// just to ensure first fs doesn't have any logic to twiddle leases
DistributedFileSystem fs2 = (DistributedFileSystem) FileSystem.newInstance(fs.getUri(), fs.getConf());
// rename the file into an existing dir
LOG.info("DMS: rename file into dir");
Path pRenamed = new Path(d, p.getName());
fs2.mkdirs(d);
fs2.rename(p, pRenamed);
Assert.assertFalse(p+" exists", fs2.exists(p));
Assert.assertTrue(pRenamed+" not found", fs2.exists(pRenamed));
Assert.assertFalse("has lease for "+p, hasLease(cluster, p));
Assert.assertTrue("no lease for "+pRenamed, hasLease(cluster, pRenamed));
Assert.assertEquals(1, leaseCount(cluster));
// rename the parent dir to a new non-existent dir
LOG.info("DMS: rename parent dir");
Path pRenamedAgain = new Path(d2, pRenamed.getName());
fs2.rename(d, d2);
// src gone
Assert.assertFalse(d+" exists", fs2.exists(d));
Assert.assertFalse("has lease for "+pRenamed, hasLease(cluster, pRenamed));
// dst checks
Assert.assertTrue(d2+" not found", fs2.exists(d2));
Assert.assertTrue(pRenamedAgain+" not found", fs2.exists(pRenamedAgain));
Assert.assertTrue("no lease for "+pRenamedAgain, hasLease(cluster, pRenamedAgain));
Assert.assertEquals(1, leaseCount(cluster));
// rename the parent dir to existing dir
// NOTE: rename w/o options moves paths into existing dir
LOG.info("DMS: rename parent again");
pRenamed = pRenamedAgain;
pRenamedAgain = new Path(new Path(d, d2.getName()), p.getName());
fs2.mkdirs(d);
fs2.rename(d2, d);
// src gone
Assert.assertFalse(d2+" exists", fs2.exists(d2));
Assert.assertFalse("no lease for "+pRenamed, hasLease(cluster, pRenamed));
// dst checks
Assert.assertTrue(d+" not found", fs2.exists(d));
Assert.assertTrue(pRenamedAgain +" not found", fs2.exists(pRenamedAgain));
Assert.assertTrue("no lease for "+pRenamedAgain, hasLease(cluster, pRenamedAgain));
Assert.assertEquals(1, leaseCount(cluster));
// rename with opts to non-existent dir
pRenamed = pRenamedAgain;
pRenamedAgain = new Path(d2, p.getName());
fs2.rename(pRenamed.getParent(), d2, Options.Rename.OVERWRITE);
// src gone
Assert.assertFalse(pRenamed.getParent() +" not found", fs2.exists(pRenamed.getParent()));
Assert.assertFalse("has lease for "+pRenamed, hasLease(cluster, pRenamed));
// dst checks
Assert.assertTrue(d2+" not found", fs2.exists(d2));
Assert.assertTrue(pRenamedAgain+" not found", fs2.exists(pRenamedAgain));
Assert.assertTrue("no lease for "+pRenamedAgain, hasLease(cluster, pRenamedAgain));
Assert.assertEquals(1, leaseCount(cluster));
// rename with opts to existing dir
// NOTE: rename with options will not move paths into the existing dir
pRenamed = pRenamedAgain;
pRenamedAgain = new Path(d, p.getName());
fs2.rename(pRenamed.getParent(), d, Options.Rename.OVERWRITE);
// src gone
Assert.assertFalse(pRenamed.getParent() +" not found", fs2.exists(pRenamed.getParent()));
Assert.assertFalse("has lease for "+pRenamed, hasLease(cluster, pRenamed));
// dst checks
Assert.assertTrue(d+" not found", fs2.exists(d));
Assert.assertTrue(pRenamedAgain+" not found", fs2.exists(pRenamedAgain));
Assert.assertTrue("no lease for "+pRenamedAgain, hasLease(cluster, pRenamedAgain));
Assert.assertEquals(1, leaseCount(cluster));
} finally {
cluster.shutdown();
}
}
@Test
public void testLease() throws Exception {
MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf).numDataNodes(2).build();

4
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestBackupNode.java
View File

@ -35,6 +35,7 @@ import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.FileUtil;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.ha.HAServiceProtocol.HAServiceState;
import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.hdfs.DFSTestUtil;
import org.apache.hadoop.hdfs.HAUtil;
@ -103,6 +104,9 @@ public class TestBackupNode {
BackupNode bn = (BackupNode)NameNode.createNameNode(
new String[]{startupOpt.getName()}, c);
assertTrue(bn.getRole() + " must be in SafeMode.", bn.isInSafeMode());
assertTrue(bn.getRole() + " must be in StandbyState",
bn.getNamesystem().getHAState()
.equalsIgnoreCase(HAServiceState.STANDBY.name()));
return bn;
}

2
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestFsLimits.java
View File

@ -158,7 +158,7 @@ public class TestFsLimits {
Class<?> generated = null;
try {
fs.verifyFsLimits(inodes, 1, child);
rootInode.addChild(child, false);
rootInode.addChild(child, false, null);
} catch (QuotaExceededException e) {
generated = e.getClass();
}

4
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestINodeFile.java
View File

@ -144,11 +144,11 @@ public class TestINodeFile {
assertEquals("f", inf.getFullPathName());
assertEquals("", inf.getLocalParentDir());
dir.addChild(inf, false);
dir.addChild(inf, false, null);
assertEquals("d"+Path.SEPARATOR+"f", inf.getFullPathName());
assertEquals("d", inf.getLocalParentDir());
root.addChild(dir, false);
root.addChild(dir, false, null);
assertEquals(Path.SEPARATOR+"d"+Path.SEPARATOR+"f", inf.getFullPathName());
assertEquals(Path.SEPARATOR+"d", dir.getFullPathName());

3
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestMetaSave.java
View File

@ -92,7 +92,8 @@ public class TestMetaSave {
DataInputStream in = new DataInputStream(fstream);
BufferedReader reader = new BufferedReader(new InputStreamReader(in));
String line = reader.readLine();
assertTrue(line.equals("3 files and directories, 2 blocks = 5 total"));
assertTrue(line.equals(
"3 files and directories, 2 blocks = 5 total filesystem objects"));
line = reader.readLine();
assertTrue(line.equals("Live Datanodes: 1"));
line = reader.readLine();

18
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestSaveNamespace.java
View File

@ -603,6 +603,24 @@ public class TestSaveNamespace {
}
}
@Test
public void testSaveNamespaceWithDanglingLease() throws Exception {
MiniDFSCluster cluster = new MiniDFSCluster.Builder(new Configuration())
.numDataNodes(1).build();
cluster.waitActive();
DistributedFileSystem fs = (DistributedFileSystem) cluster.getFileSystem();
try {
cluster.getNamesystem().leaseManager.addLease("me", "/non-existent");
fs.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
cluster.getNameNodeRpc().saveNamespace();
fs.setSafeMode(SafeModeAction.SAFEMODE_LEAVE);
} finally {
if (cluster != null) {
cluster.shutdown();
}
}
}
private void doAnEdit(FSNamesystem fsn, int id) throws IOException {
// Make an edit
fsn.mkdirs(

8
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestSnapshotPathINodes.java
View File

@ -430,10 +430,10 @@ public class TestSnapshotPathINodes {
// The number of inodes should be equal to components.length
assertEquals(newInodes.length, components.length);
// The last INode should be associated with file1
assertEquals(newInodes[components.length - 1].getFullPathName(),
file1.toString());
final int last = components.length - 1;
assertEquals(newInodes[last].getFullPathName(), file1.toString());
// The modification time of the INode for file3 should have been changed
Assert.assertFalse(inodes[components.length - 1].getModificationTime() ==
newInodes[components.length - 1].getModificationTime());
Assert.assertFalse(inodes[last].getModificationTime()
== newInodes[last].getModificationTime());
}
}

2
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/snapshot/TestINodeDirectoryWithSnapshot.java
View File

@ -221,7 +221,7 @@ public class TestINodeDirectoryWithSnapshot {
Assert.assertTrue(i >= 0);
final INodeDirectory oldinode = (INodeDirectory)current.get(i);
final INodeDirectory newinode = new INodeDirectory(oldinode);
newinode.setModificationTime(oldinode.getModificationTime() + 1);
newinode.updateModificationTime(oldinode.getModificationTime() + 1);
current.set(i, newinode);
if (diff != null) {

3
hadoop-mapreduce-project/CHANGES.txt
View File

@ -604,6 +604,9 @@ Release 0.23.6 - UNRELEASED
MAPREDUCE-4817. Hardcoded task ping timeout kills tasks localizing large
amounts of data (tgraves)
MAPREDUCE-4836. Elapsed time for running tasks on AM web UI tasks page is 0
(Ravi Prakash via jeagles)
Release 0.23.5 - UNRELEASED
INCOMPATIBLE CHANGES

5
hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/webapp/dao/TaskInfo.java
View File

@ -59,11 +59,12 @@ public class TaskInfo {
TaskReport report = task.getReport();
this.startTime = report.getStartTime();
this.finishTime = report.getFinishTime();
this.elapsedTime = Times.elapsed(this.startTime, this.finishTime, false);
this.state = report.getTaskState();
this.elapsedTime = Times.elapsed(this.startTime, this.finishTime,
this.state == TaskState.RUNNING);
if (this.elapsedTime == -1) {
this.elapsedTime = 0;
}
this.state = report.getTaskState();
this.progress = report.getProgress() * 100;
this.id = MRApps.toString(task.getID());
this.taskNum = task.getID().getId();

6
hadoop-yarn-project/CHANGES.txt
View File

@ -117,7 +117,11 @@ Release 2.0.3-alpha - Unreleased
YARN-229. Remove old unused RM recovery code. (Bikas Saha via acmurthy)
YARN-187. Add hierarchical queues to the fair scheduler. (Sandy Ryza via tomwhite)
YARN-187. Add hierarchical queues to the fair scheduler.
(Sandy Ryza via tomwhite)
YARN-72. NM should handle cleaning up containers when it shuts down.
(Sandy Ryza via tomwhite)
Release 2.0.2-alpha - 2012-09-07

12
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/CMgrCompletedContainersEvent.java
View File

@ -25,13 +25,23 @@ import org.apache.hadoop.yarn.api.records.ContainerId;
public class CMgrCompletedContainersEvent extends ContainerManagerEvent {
private List<ContainerId> containerToCleanup;
private Reason reason;
public CMgrCompletedContainersEvent(List<ContainerId> containersToCleanup) {
public CMgrCompletedContainersEvent(List<ContainerId> containersToCleanup, Reason reason) {
super(ContainerManagerEventType.FINISH_CONTAINERS);
this.containerToCleanup = containersToCleanup;
this.reason = reason;
}
public List<ContainerId> getContainersToCleanup() {
return this.containerToCleanup;
}
public Reason getReason() {
return reason;
}
public static enum Reason {
ON_SHUTDOWN, BY_RESOURCEMANAGER
}
}

66
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java
View File

@ -19,6 +19,9 @@
package org.apache.hadoop.yarn.server.nodemanager;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.ConcurrentSkipListMap;
@ -62,13 +65,23 @@ public class NodeManager extends CompositeService implements
*/
public static final int SHUTDOWN_HOOK_PRIORITY = 30;
/**
* Extra duration to wait for containers to be killed on shutdown.
*/
private static final int SHUTDOWN_CLEANUP_SLOP_MS = 1000;
private static final Log LOG = LogFactory.getLog(NodeManager.class);
protected final NodeManagerMetrics metrics = NodeManagerMetrics.create();
private ApplicationACLsManager aclsManager;
private NodeHealthCheckerService nodeHealthChecker;
private LocalDirsHandlerService dirsHandler;
private Context context;
private AsyncDispatcher dispatcher;
private ContainerManagerImpl containerManager;
private static CompositeServiceShutdownHook nodeManagerShutdownHook;
private long waitForContainersOnShutdownMillis;
public NodeManager() {
super(NodeManager.class.getName());
}
@ -115,7 +128,7 @@ public class NodeManager extends CompositeService implements
containerTokenSecretManager = new NMContainerTokenSecretManager(conf);
}
Context context = new NMContext(containerTokenSecretManager);
this.context = new NMContext(containerTokenSecretManager);
this.aclsManager = new ApplicationACLsManager(conf);
@ -131,7 +144,7 @@ public class NodeManager extends CompositeService implements
addService(del);
// NodeManager level dispatcher
AsyncDispatcher dispatcher = new AsyncDispatcher();
this.dispatcher = new AsyncDispatcher();
nodeHealthChecker = new NodeHealthCheckerService();
addService(nodeHealthChecker);
@ -144,7 +157,7 @@ public class NodeManager extends CompositeService implements
NodeResourceMonitor nodeResourceMonitor = createNodeResourceMonitor();
addService(nodeResourceMonitor);
ContainerManagerImpl containerManager =
containerManager =
createContainerManager(context, exec, del, nodeStatusUpdater,
this.aclsManager, dirsHandler);
addService(containerManager);
@ -162,6 +175,13 @@ public class NodeManager extends CompositeService implements
// so that we make sure everything is up before registering with RM.
addService(nodeStatusUpdater);
waitForContainersOnShutdownMillis =
conf.getLong(YarnConfiguration.NM_SLEEP_DELAY_BEFORE_SIGKILL_MS,
YarnConfiguration.DEFAULT_NM_SLEEP_DELAY_BEFORE_SIGKILL_MS) +
conf.getLong(YarnConfiguration.NM_PROCESS_KILL_WAIT_MS,
YarnConfiguration.DEFAULT_NM_PROCESS_KILL_WAIT_MS) +
SHUTDOWN_CLEANUP_SLOP_MS;
super.init(conf);
// TODO add local dirs to del
}
@ -178,10 +198,45 @@ public class NodeManager extends CompositeService implements
@Override
public void stop() {
cleanupContainers();
super.stop();
DefaultMetricsSystem.shutdown();
}
@SuppressWarnings("unchecked")
private void cleanupContainers() {
Map<ContainerId, Container> containers = context.getContainers();
if (containers.isEmpty()) {
return;
}
LOG.info("Containers still running on shutdown: " + containers.keySet());
List<ContainerId> containerIds = new ArrayList<ContainerId>(containers.keySet());
dispatcher.getEventHandler().handle(
new CMgrCompletedContainersEvent(containerIds,
CMgrCompletedContainersEvent.Reason.ON_SHUTDOWN));
LOG.info("Waiting for containers to be killed");
long waitStartTime = System.currentTimeMillis();
while (!containers.isEmpty() &&
System.currentTimeMillis() - waitStartTime < waitForContainersOnShutdownMillis) {
try {
Thread.sleep(1000);
} catch (InterruptedException ex) {
LOG.warn("Interrupted while sleeping on container kill", ex);
}
}
// All containers killed
if (containers.isEmpty()) {
LOG.info("All containers in DONE state");
} else {
LOG.info("Done waiting for containers to be killed. Still alive: " +
containers.keySet());
}
}
public static class NMContext implements Context {
private final NodeId nodeId = Records.newRecord(NodeId.class);
@ -283,6 +338,11 @@ public class NodeManager extends CompositeService implements
return new NodeManager();
}
// For testing
ContainerManagerImpl getContainerManager() {
return containerManager;
}
public static void main(String[] args) {
Thread.setDefaultUncaughtExceptionHandler(new YarnUncaughtExceptionHandler());
StringUtils.startupShutdownMessage(NodeManager.class, args, LOG);

3
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeStatusUpdaterImpl.java
View File

@ -363,7 +363,8 @@ public class NodeStatusUpdaterImpl extends AbstractService implements
.getContainersToCleanupList();
if (containersToCleanup.size() != 0) {
dispatcher.getEventHandler().handle(
new CMgrCompletedContainersEvent(containersToCleanup));
new CMgrCompletedContainersEvent(containersToCleanup,
CMgrCompletedContainersEvent.Reason.BY_RESOURCEMANAGER));
}
List<ApplicationId> appsToCleanup =
response.getApplicationsToCleanupList();

13
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/ContainerManagerImpl.java
View File

@ -23,6 +23,8 @@ import static org.apache.hadoop.yarn.service.Service.STATE.STARTED;
import java.io.IOException;
import java.net.InetSocketAddress;
import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.Set;
@ -593,9 +595,16 @@ public class ContainerManagerImpl extends CompositeService implements
(CMgrCompletedContainersEvent) event;
for (ContainerId container : containersFinishedEvent
.getContainersToCleanup()) {
String diagnostic = "";
if (containersFinishedEvent.getReason() ==
CMgrCompletedContainersEvent.Reason.ON_SHUTDOWN) {
diagnostic = "Container Killed on Shutdown";
} else if (containersFinishedEvent.getReason() ==
CMgrCompletedContainersEvent.Reason.BY_RESOURCEMANAGER) {
diagnostic = "Container Killed by ResourceManager";
}
this.dispatcher.getEventHandler().handle(
new ContainerKillEvent(container,
"Container Killed by ResourceManager"));
new ContainerKillEvent(container, diagnostic));
}
break;
default:

92
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/MockNodeStatusUpdater.java Normal file
View File

@ -0,0 +1,92 @@
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.yarn.server.nodemanager;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.yarn.event.Dispatcher;
import org.apache.hadoop.yarn.exceptions.YarnRemoteException;
import org.apache.hadoop.yarn.factories.RecordFactory;
import org.apache.hadoop.yarn.factory.providers.RecordFactoryProvider;
import org.apache.hadoop.yarn.server.api.ResourceTracker;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatRequest;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatResponse;
import org.apache.hadoop.yarn.server.api.protocolrecords.RegisterNodeManagerRequest;
import org.apache.hadoop.yarn.server.api.protocolrecords.RegisterNodeManagerResponse;
import org.apache.hadoop.yarn.server.api.records.HeartbeatResponse;
import org.apache.hadoop.yarn.server.api.records.NodeStatus;
import org.apache.hadoop.yarn.server.api.records.RegistrationResponse;
import org.apache.hadoop.yarn.server.nodemanager.metrics.NodeManagerMetrics;
/**
* This class allows a node manager to run without without communicating with a
* real RM.
*/
public class MockNodeStatusUpdater extends NodeStatusUpdaterImpl {
static final Log LOG = LogFactory.getLog(MockNodeStatusUpdater.class);
private static final RecordFactory recordFactory = RecordFactoryProvider
.getRecordFactory(null);
private ResourceTracker resourceTracker;
public MockNodeStatusUpdater(Context context, Dispatcher dispatcher,
NodeHealthCheckerService healthChecker, NodeManagerMetrics metrics) {
super(context, dispatcher, healthChecker, metrics);
resourceTracker = new MockResourceTracker();
}
@Override
protected ResourceTracker getRMClient() {
return resourceTracker;
}
private static class MockResourceTracker implements ResourceTracker {
private int heartBeatID;
@Override
public RegisterNodeManagerResponse registerNodeManager(
RegisterNodeManagerRequest request) throws YarnRemoteException {
RegistrationResponse regResponse = recordFactory
.newRecordInstance(RegistrationResponse.class);
RegisterNodeManagerResponse response = recordFactory
.newRecordInstance(RegisterNodeManagerResponse.class);
response.setRegistrationResponse(regResponse);
return response;
}
@Override
public NodeHeartbeatResponse nodeHeartbeat(NodeHeartbeatRequest request)
throws YarnRemoteException {
NodeStatus nodeStatus = request.getNodeStatus();
LOG.info("Got heartbeat number " + heartBeatID);
nodeStatus.setResponseId(heartBeatID++);
HeartbeatResponse response = recordFactory
.newRecordInstance(HeartbeatResponse.class);
response.setResponseId(heartBeatID);
NodeHeartbeatResponse nhResponse = recordFactory
.newRecordInstance(NodeHeartbeatResponse.class);
nhResponse.setHeartbeatResponse(response);
return nhResponse;
}
}
}

222
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/TestNodeManagerShutdown.java Normal file
View File

@ -0,0 +1,222 @@
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.yarn.server.nodemanager;
import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.File;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import junit.framework.Assert;
import org.apache.hadoop.fs.FileContext;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.UnsupportedFileSystemException;
import org.apache.hadoop.yarn.api.protocolrecords.GetContainerStatusRequest;
import org.apache.hadoop.yarn.api.protocolrecords.StartContainerRequest;
import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
import org.apache.hadoop.yarn.api.records.ContainerId;
import org.apache.hadoop.yarn.api.records.ContainerLaunchContext;
import org.apache.hadoop.yarn.api.records.ContainerState;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.LocalResource;
import org.apache.hadoop.yarn.api.records.LocalResourceType;
import org.apache.hadoop.yarn.api.records.LocalResourceVisibility;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.URL;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.event.Dispatcher;
import org.apache.hadoop.yarn.factories.RecordFactory;
import org.apache.hadoop.yarn.factory.providers.RecordFactoryProvider;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.ContainerManagerImpl;
import org.apache.hadoop.yarn.util.ConverterUtils;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
public class TestNodeManagerShutdown {
static final File basedir =
new File("target", TestNodeManagerShutdown.class.getName());
static final File tmpDir = new File(basedir, "tmpDir");
static final File logsDir = new File(basedir, "logs");
static final File remoteLogsDir = new File(basedir, "remotelogs");
static final File nmLocalDir = new File(basedir, "nm0");
static final File processStartFile = new File(tmpDir, "start_file.txt")
.getAbsoluteFile();
static final RecordFactory recordFactory = RecordFactoryProvider
.getRecordFactory(null);
static final String user = "nobody";
private FileContext localFS;
@Before
public void setup() throws UnsupportedFileSystemException {
localFS = FileContext.getLocalFSFileContext();
tmpDir.mkdirs();
logsDir.mkdirs();
remoteLogsDir.mkdirs();
nmLocalDir.mkdirs();
}
@After
public void tearDown() throws IOException, InterruptedException {
localFS.delete(new Path(basedir.getPath()), true);
}
@Test
public void testKillContainersOnShutdown() throws IOException {
NodeManager nm = getNodeManager();
nm.init(createNMConfig());
nm.start();
ContainerManagerImpl containerManager = nm.getContainerManager();
File scriptFile = createUnhaltingScriptFile();
ContainerLaunchContext containerLaunchContext =
recordFactory.newRecordInstance(ContainerLaunchContext.class);
// Construct the Container-id
ContainerId cId = createContainerId();
containerLaunchContext.setContainerId(cId);
containerLaunchContext.setUser(user);
URL localResourceUri =
ConverterUtils.getYarnUrlFromPath(localFS
.makeQualified(new Path(scriptFile.getAbsolutePath())));
LocalResource localResource =
recordFactory.newRecordInstance(LocalResource.class);
localResource.setResource(localResourceUri);
localResource.setSize(-1);
localResource.setVisibility(LocalResourceVisibility.APPLICATION);
localResource.setType(LocalResourceType.FILE);
localResource.setTimestamp(scriptFile.lastModified());
String destinationFile = "dest_file";
Map<String, LocalResource> localResources =
new HashMap<String, LocalResource>();
localResources.put(destinationFile, localResource);
containerLaunchContext.setLocalResources(localResources);
containerLaunchContext.setUser(containerLaunchContext.getUser());
List<String> commands = new ArrayList<String>();
commands.add("/bin/bash");
commands.add(scriptFile.getAbsolutePath());
containerLaunchContext.setCommands(commands);
containerLaunchContext.setResource(recordFactory
.newRecordInstance(Resource.class));
containerLaunchContext.getResource().setMemory(1024);
StartContainerRequest startRequest = recordFactory.newRecordInstance(StartContainerRequest.class);
startRequest.setContainerLaunchContext(containerLaunchContext);
containerManager.startContainer(startRequest);
GetContainerStatusRequest request =
recordFactory.newRecordInstance(GetContainerStatusRequest.class);
request.setContainerId(cId);
ContainerStatus containerStatus =
containerManager.getContainerStatus(request).getStatus();
Assert.assertEquals(ContainerState.RUNNING, containerStatus.getState());
try {Thread.sleep(5000);} catch (InterruptedException ex) {ex.printStackTrace();}
nm.stop();
// Now verify the contents of the file
// Script generates a message when it receives a sigterm
// so we look for that
BufferedReader reader =
new BufferedReader(new FileReader(processStartFile));
boolean foundSigTermMessage = false;
while (true) {
String line = reader.readLine();
if (line == null) {
break;
}
if (line.contains("SIGTERM")) {
foundSigTermMessage = true;
break;
}
}
Assert.assertTrue("Did not find sigterm message", foundSigTermMessage);
reader.close();
}
private ContainerId createContainerId() {
ApplicationId appId = recordFactory.newRecordInstance(ApplicationId.class);
appId.setClusterTimestamp(0);
appId.setId(0);
ApplicationAttemptId appAttemptId =
recordFactory.newRecordInstance(ApplicationAttemptId.class);
appAttemptId.setApplicationId(appId);
appAttemptId.setAttemptId(1);
ContainerId containerId =
recordFactory.newRecordInstance(ContainerId.class);
containerId.setApplicationAttemptId(appAttemptId);
return containerId;
}
private YarnConfiguration createNMConfig() {
YarnConfiguration conf = new YarnConfiguration();
conf.setInt(YarnConfiguration.NM_PMEM_MB, 5*1024); // 5GB
conf.set(YarnConfiguration.NM_ADDRESS, "127.0.0.1:12345");
conf.set(YarnConfiguration.NM_LOCALIZER_ADDRESS, "127.0.0.1:12346");
conf.set(YarnConfiguration.NM_LOG_DIRS, logsDir.getAbsolutePath());
conf.set(YarnConfiguration.NM_REMOTE_APP_LOG_DIR, remoteLogsDir.getAbsolutePath());
conf.set(YarnConfiguration.NM_LOCAL_DIRS, nmLocalDir.getAbsolutePath());
return conf;
}
/**
* Creates a script to run a container that will run forever unless
* stopped by external means.
*/
private File createUnhaltingScriptFile() throws IOException {
File scriptFile = new File(tmpDir, "scriptFile.sh");
BufferedWriter fileWriter = new BufferedWriter(new FileWriter(scriptFile));
fileWriter.write("#!/bin/bash\n\n");
fileWriter.write("echo \"Running testscript for delayed kill\"\n");
fileWriter.write("hello=\"Got SIGTERM\"\n");
fileWriter.write("umask 0\n");
fileWriter.write("trap \"echo $hello >> " + processStartFile + "\" SIGTERM\n");
fileWriter.write("echo \"Writing pid to start file\"\n");
fileWriter.write("echo $$ >> " + processStartFile + "\n");
fileWriter.write("while true; do\nsleep 1s;\ndone\n");
fileWriter.close();
return scriptFile;
}
private NodeManager getNodeManager() {
return new NodeManager() {
@Override
protected NodeStatusUpdater createNodeStatusUpdater(Context context,
Dispatcher dispatcher, NodeHealthCheckerService healthChecker) {
MockNodeStatusUpdater myNodeStatusUpdater = new MockNodeStatusUpdater(
context, dispatcher, healthChecker, metrics);
return myNodeStatusUpdater;
}
};
}
}