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

HDFS-6911. Archival Storage: check if a block is already scheduled in Mover. Contributed by Tsz Wo Nicholas Sze.

This commit is contained in:
Jing Zhao 2014-08-27 10:38:10 -07:00
parent 555900a9dc
commit 8ea20b53a8
5 changed files with 228 additions and 66 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Dispatcher.java
View File

@ -175,11 +175,7 @@ public class Dispatcher {
private DDatanode proxySource;
private StorageGroup target;
private PendingMove() {
}
public PendingMove(DBlock block, Source source, StorageGroup target) {
this.block = block;
private PendingMove(Source source, StorageGroup target) {
this.source = source;
this.target = target;
}
@ -199,9 +195,11 @@ public class Dispatcher {
* @return true if a block and its proxy are chosen; false otherwise
*/
private boolean chooseBlockAndProxy() {
// source and target must have the same storage type
final StorageType t = source.getStorageType();
// iterate all source's blocks until find a good one
for (Iterator<DBlock> i = source.getBlockIterator(); i.hasNext();) {
if (markMovedIfGoodBlock(i.next())) {
if (markMovedIfGoodBlock(i.next(), t)) {
i.remove();
return true;
}
@ -212,10 +210,10 @@ public class Dispatcher {
/**
* @return true if the given block is good for the tentative move.
*/
private boolean markMovedIfGoodBlock(DBlock block) {
private boolean markMovedIfGoodBlock(DBlock block, StorageType targetStorageType) {
synchronized (block) {
synchronized (movedBlocks) {
if (isGoodBlockCandidate(source, target, block)) {
if (isGoodBlockCandidate(source, target, targetStorageType, block)) {
this.block = block;
if (chooseProxySource()) {
movedBlocks.put(block);
@ -235,7 +233,7 @@ public class Dispatcher {
*
* @return true if a proxy is found; otherwise false
*/
public boolean chooseProxySource() {
private boolean chooseProxySource() {
final DatanodeInfo targetDN = target.getDatanodeInfo();
// if node group is supported, first try add nodes in the same node group
if (cluster.isNodeGroupAware()) {
@ -440,6 +438,18 @@ public class Dispatcher {
scheduledSize = 0L;
}
private PendingMove addPendingMove(DBlock block, final PendingMove pm) {
if (getDDatanode().addPendingBlock(pm)) {
if (pm.markMovedIfGoodBlock(block, getStorageType())) {
incScheduledSize(pm.block.getNumBytes());
return pm;
} else {
getDDatanode().removePendingBlock(pm);
}
}
return null;
}
/** @return the name for display */
String getDisplayName() {
return datanode + ":" + storageType;
@ -599,8 +609,11 @@ public class Dispatcher {
/** Decide if the given block is a good candidate to move or not */
private boolean isGoodBlockCandidate(DBlock block) {
// source and target must have the same storage type
final StorageType sourceStorageType = getStorageType();
for (Task t : tasks) {
if (Dispatcher.this.isGoodBlockCandidate(this, t.target, block)) {
if (Dispatcher.this.isGoodBlockCandidate(this, t.target,
sourceStorageType, block)) {
return true;
}
}
@ -620,11 +633,9 @@ public class Dispatcher {
for (Iterator<Task> i = tasks.iterator(); i.hasNext();) {
final Task task = i.next();
final DDatanode target = task.target.getDDatanode();
PendingMove pendingBlock = new PendingMove();
final PendingMove pendingBlock = new PendingMove(this, task.target);
if (target.addPendingBlock(pendingBlock)) {
// target is not busy, so do a tentative block allocation
pendingBlock.source = this;
pendingBlock.target = task.target;
if (pendingBlock.chooseBlockAndProxy()) {
long blockSize = pendingBlock.block.getNumBytes();
incScheduledSize(-blockSize);
@ -641,6 +652,11 @@ public class Dispatcher {
}
return null;
}
/** Add a pending move */
public PendingMove addPendingMove(DBlock block, StorageGroup target) {
return target.addPendingMove(block, new PendingMove(this, target));
}
/** Iterate all source's blocks to remove moved ones */
private void removeMovedBlocks() {
@ -901,12 +917,6 @@ public class Dispatcher {
}
}
private boolean isGoodBlockCandidate(StorageGroup source, StorageGroup target,
DBlock block) {
// match source and target storage type
return isGoodBlockCandidate(source, target, source.getStorageType(), block);
}
/**
* Decide if the block is a good candidate to be moved from source to target.
* A block is a good candidate if
@ -914,7 +924,7 @@ public class Dispatcher {
* 2. the block does not have a replica on the target;
* 3. doing the move does not reduce the number of racks that the block has
*/
public boolean isGoodBlockCandidate(StorageGroup source, StorageGroup target,
private boolean isGoodBlockCandidate(StorageGroup source, StorageGroup target,
StorageType targetStorageType, DBlock block) {
if (target.storageType != targetStorageType) {
return false;

15
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Matcher.java
View File

@ -31,6 +31,11 @@ public interface Matcher {
public boolean match(NetworkTopology cluster, Node left, Node right) {
return cluster.isOnSameNodeGroup(left, right);
}
@Override
public String toString() {
return "SAME_NODE_GROUP";
}
};
/** Match datanodes in the same rack. */
@ -39,6 +44,11 @@ public interface Matcher {
public boolean match(NetworkTopology cluster, Node left, Node right) {
return cluster.isOnSameRack(left, right);
}
@Override
public String toString() {
return "SAME_RACK";
}
};
/** Match any datanode with any other datanode. */
@ -47,5 +57,10 @@ public interface Matcher {
public boolean match(NetworkTopology cluster, Node left, Node right) {
return left != right;
}
@Override
public String toString() {
return "ANY_OTHER";
}
};
}

65
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/mover/Mover.java
View File

@ -50,7 +50,7 @@ import java.util.*;
public class Mover {
static final Log LOG = LogFactory.getLog(Mover.class);
private static final Path MOVER_ID_PATH = new Path("/system/mover.id");
static final Path MOVER_ID_PATH = new Path("/system/mover.id");
private static class StorageMap {
private final StorageGroupMap<Source> sources
@ -111,22 +111,25 @@ public class Mover {
this.storages = new StorageMap();
this.blockStoragePolicies = BlockStoragePolicy.readBlockStorageSuite(conf);
}
private ExitStatus run() {
try {
final List<DatanodeStorageReport> reports = dispatcher.init();
for(DatanodeStorageReport r : reports) {
final DDatanode dn = dispatcher.newDatanode(r.getDatanodeInfo());
for(StorageType t : StorageType.asList()) {
final long maxRemaining = getMaxRemaining(r, t);
if (maxRemaining > 0L) {
final Source source = dn.addSource(t, Long.MAX_VALUE, dispatcher);
final StorageGroup target = dn.addTarget(t, maxRemaining);
storages.add(source, target);
}
void init() throws IOException {
final List<DatanodeStorageReport> reports = dispatcher.init();
for(DatanodeStorageReport r : reports) {
final DDatanode dn = dispatcher.newDatanode(r.getDatanodeInfo());
for(StorageType t : StorageType.asList()) {
final long maxRemaining = getMaxRemaining(r, t);
if (maxRemaining > 0L) {
final Source source = dn.addSource(t, Long.MAX_VALUE, dispatcher);
final StorageGroup target = dn.addTarget(t, maxRemaining);
storages.add(source, target);
}
}
}
}
private ExitStatus run() {
try {
init();
new Processor().processNamespace();
return ExitStatus.IN_PROGRESS;
@ -141,6 +144,14 @@ public class Mover {
}
}
DBlock newDBlock(Block block, List<MLocation> locations) {
final DBlock db = new DBlock(block);
for(MLocation ml : locations) {
db.addLocation(storages.getTarget(ml));
}
return db;
}
private static long getMaxRemaining(DatanodeStorageReport report, StorageType t) {
long max = 0L;
for(StorageReport r : report.getStorageReports()) {
@ -169,11 +180,11 @@ public class Mover {
return sb.toString();
}
private class Processor {
class Processor {
private final DFSClient dfs;
private final List<String> snapshottableDirs = new ArrayList<String>();
private Processor() {
Processor() {
dfs = dispatcher.getDistributedFileSystem().getClient();
}
@ -290,15 +301,11 @@ public class Mover {
}
}
}
void scheduleMoves4Block(StorageTypeDiff diff, LocatedBlock lb) {
final List<MLocation> locations = MLocation.toLocations(lb);
Collections.shuffle(locations);
final DBlock db = new DBlock(lb.getBlock().getLocalBlock());
for(MLocation ml : locations) {
db.addLocation(storages.getTarget(ml));
}
final DBlock db = newDBlock(lb.getBlock().getLocalBlock(), locations);
for(final Iterator<StorageType> i = diff.existing.iterator(); i.hasNext(); ) {
final StorageType t = i.next();
@ -310,12 +317,18 @@ public class Mover {
if (scheduleMoveReplica(db, ml, source, diff.expected)) {
i.remove();
j.remove();
return;
}
}
}
}
}
boolean scheduleMoveReplica(DBlock db, MLocation ml,
List<StorageType> targetTypes) {
return scheduleMoveReplica(db, ml, storages.getSource(ml), targetTypes);
}
boolean scheduleMoveReplica(DBlock db, MLocation ml, Source source,
List<StorageType> targetTypes) {
if (dispatcher.getCluster().isNodeGroupAware()) {
@ -341,12 +354,10 @@ public class Mover {
for(final Iterator<StorageType> i = targetTypes.iterator(); i.hasNext(); ) {
final StorageType t = i.next();
for(StorageGroup target : storages.getTargetStorages(t)) {
if (matcher.match(cluster, ml.datanode, target.getDatanodeInfo())
&& dispatcher.isGoodBlockCandidate(source, target, t, db)) {
final PendingMove pm = dispatcher.new PendingMove(db, source, target);
if (pm.chooseProxySource()) {
if (matcher.match(cluster, ml.datanode, target.getDatanodeInfo())) {
final PendingMove pm = source.addPendingMove(db, target);
if (pm != null) {
i.remove();
target.incScheduledSize(ml.size);
dispatcher.executePendingMove(pm);
return true;
}

79
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestBlockStoragePolicy.java
View File

@ -20,7 +20,6 @@ package org.apache.hadoop.hdfs;
import static org.apache.hadoop.hdfs.BlockStoragePolicy.ID_UNSPECIFIED;
import java.io.FileNotFoundException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.EnumSet;
import java.util.HashMap;
@ -62,20 +61,6 @@ public class TestBlockStoragePolicy {
static final byte WARM = (byte) 8;
static final byte HOT = (byte) 12;
static final List<List<StorageType>> chosens = new ArrayList<List<StorageType>>();
static {
chosens.add(Arrays.<StorageType>asList());
chosens.add(Arrays.asList(StorageType.DISK));
chosens.add(Arrays.asList(StorageType.ARCHIVE));
chosens.add(Arrays.asList(StorageType.DISK, StorageType.DISK));
chosens.add(Arrays.asList(StorageType.DISK, StorageType.ARCHIVE));
chosens.add(Arrays.asList(StorageType.ARCHIVE, StorageType.ARCHIVE));
chosens.add(Arrays.asList(StorageType.DISK, StorageType.DISK, StorageType.DISK));
chosens.add(Arrays.asList(StorageType.DISK, StorageType.DISK, StorageType.ARCHIVE));
chosens.add(Arrays.asList(StorageType.DISK, StorageType.ARCHIVE, StorageType.ARCHIVE));
chosens.add(Arrays.asList(StorageType.ARCHIVE, StorageType.ARCHIVE, StorageType.ARCHIVE));
}
@Test
public void testDefaultPolicies() {
final Map<Byte, String> expectedPolicyStrings = new HashMap<Byte, String>();
@ -126,6 +111,17 @@ public class TestBlockStoragePolicy {
}
}
static StorageType[] newStorageTypes(int nDisk, int nArchive) {
final StorageType[] t = new StorageType[nDisk + nArchive];
Arrays.fill(t, 0, nDisk, StorageType.DISK);
Arrays.fill(t, nDisk, t.length, StorageType.ARCHIVE);
return t;
}
static List<StorageType> asList(int nDisk, int nArchive) {
return Arrays.asList(newStorageTypes(nDisk, nArchive));
}
static void assertStorageType(List<StorageType> computed, short replication,
StorageType... answers) {
Assert.assertEquals(replication, computed.size());
@ -369,10 +365,14 @@ public class TestBlockStoragePolicy {
final BlockStoragePolicy cold = POLICY_SUITE.getPolicy(COLD);
final short replication = 3;
for(List<StorageType> c : chosens) {
method.checkChooseStorageTypes(hot, replication, c);
method.checkChooseStorageTypes(warm, replication, c);
method.checkChooseStorageTypes(cold, replication, c);
for(int n = 0; n <= 3; n++) {
for(int d = 0; d <= n; d++) {
final int a = n - d;
final List<StorageType> chosen = asList(d, a);
method.checkChooseStorageTypes(hot, replication, chosen);
method.checkChooseStorageTypes(warm, replication, chosen);
method.checkChooseStorageTypes(cold, replication, chosen);
}
}
}
@ -714,6 +714,47 @@ public class TestBlockStoragePolicy {
Assert.assertArrayEquals(expected, computed);
}
@Test
public void testChooseExcess() {
final BlockStoragePolicy hot = POLICY_SUITE.getPolicy(HOT);
final BlockStoragePolicy warm = POLICY_SUITE.getPolicy(WARM);
final BlockStoragePolicy cold = POLICY_SUITE.getPolicy(COLD);
final short replication = 3;
for(int n = 0; n <= 6; n++) {
for(int d = 0; d <= n; d++) {
final int a = n - d;
final List<StorageType> chosen = asList(d, a);
{
final int nDisk = Math.max(0, d - replication);
final int nArchive = a;
final StorageType[] expected = newStorageTypes(nDisk, nArchive);
checkChooseExcess(hot, replication, chosen, expected);
}
{
final int nDisk = Math.max(0, d - 1);
final int nArchive = Math.max(0, a - replication + 1);
final StorageType[] expected = newStorageTypes(nDisk, nArchive);
checkChooseExcess(warm, replication, chosen, expected);
}
{
final int nDisk = d;
final int nArchive = Math.max(0, a - replication );
final StorageType[] expected = newStorageTypes(nDisk, nArchive);
checkChooseExcess(cold, replication, chosen, expected);
}
}
}
}
static void checkChooseExcess(BlockStoragePolicy p, short replication,
List<StorageType> chosen, StorageType... expected) {
final List<StorageType> types = p.chooseExcess(replication, chosen);
assertStorageTypes(types, expected);
}
private void checkDirectoryListing(HdfsFileStatus[] stats, byte... policies) {
Assert.assertEquals(stats.length, policies.length);
for (int i = 0; i < stats.length; i++) {

85
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/mover/TestMover.java Normal file
View File

@ -0,0 +1,85 @@
/**
* 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.hdfs.server.mover;
import java.io.IOException;
import java.net.URI;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.List;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hdfs.DFSUtil;
import org.apache.hadoop.hdfs.DistributedFileSystem;
import org.apache.hadoop.hdfs.HdfsConfiguration;
import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.StorageType;
import org.apache.hadoop.hdfs.protocol.LocatedBlock;
import org.apache.hadoop.hdfs.server.balancer.Dispatcher.DBlock;
import org.apache.hadoop.hdfs.server.balancer.NameNodeConnector;
import org.apache.hadoop.hdfs.server.mover.Mover.MLocation;
import org.junit.Assert;
import org.junit.Test;
public class TestMover {
static Mover newMover(Configuration conf) throws IOException {
final Collection<URI> namenodes = DFSUtil.getNsServiceRpcUris(conf);
Assert.assertEquals(1, namenodes.size());
final List<NameNodeConnector> nncs = NameNodeConnector.newNameNodeConnectors(
namenodes, Mover.class.getSimpleName(), Mover.MOVER_ID_PATH, conf);
return new Mover(nncs.get(0), conf);
}
@Test
public void testScheduleSameBlock() throws IOException {
final Configuration conf = new HdfsConfiguration();
final MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf)
.numDataNodes(4).build();
try {
cluster.waitActive();
final DistributedFileSystem dfs = cluster.getFileSystem();
final String file = "/testScheduleSameBlock/file";
{
final FSDataOutputStream out = dfs.create(new Path(file));
out.writeChars("testScheduleSameBlock");
out.close();
}
final Mover mover = newMover(conf);
mover.init();
final Mover.Processor processor = mover.new Processor();
final LocatedBlock lb = dfs.getClient().getLocatedBlocks(file, 0).get(0);
final List<MLocation> locations = MLocation.toLocations(lb);
final MLocation ml = locations.get(0);
final DBlock db = mover.newDBlock(lb.getBlock().getLocalBlock(), locations);
final List<StorageType> storageTypes = new ArrayList<StorageType>(
Arrays.asList(StorageType.DEFAULT, StorageType.DEFAULT));
Assert.assertTrue(processor.scheduleMoveReplica(db, ml, storageTypes));
Assert.assertFalse(processor.scheduleMoveReplica(db, ml, storageTypes));
} finally {
cluster.shutdown();
}
}
}