HDFS-10843. Update space quota when a UC block is completed rather than committed. Contributed by Erik Krogen.

This commit is contained in:
Konstantin V Shvachko 2016-09-22 13:57:37 -07:00
parent bbdf350ff9
commit a5bb88c8e0
6 changed files with 313 additions and 127 deletions

View File

@ -83,6 +83,7 @@ import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.BlockUCState;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.ReplicaState;
import org.apache.hadoop.hdfs.server.namenode.CachedBlock;
import org.apache.hadoop.hdfs.server.namenode.INode.BlocksMapUpdateInfo;
import org.apache.hadoop.hdfs.server.namenode.INodesInPath;
import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.namenode.Namesystem;
import org.apache.hadoop.hdfs.server.namenode.ha.HAContext;
@ -786,12 +787,13 @@ public class BlockManager implements BlockStatsMXBean {
*
* @param bc block collection
* @param commitBlock - contains client reported block length and generation
* @param iip - INodes in path to bc
* @return true if the last block is changed to committed state.
* @throws IOException if the block does not have at least a minimal number
* of replicas reported from data-nodes.
*/
public boolean commitOrCompleteLastBlock(BlockCollection bc,
Block commitBlock) throws IOException {
Block commitBlock, INodesInPath iip) throws IOException {
if(commitBlock == null)
return false; // not committing, this is a block allocation retry
BlockInfo lastBlock = bc.getLastBlock();
@ -811,7 +813,7 @@ public class BlockManager implements BlockStatsMXBean {
if (committed) {
addExpectedReplicasToPending(lastBlock);
}
completeBlock(lastBlock, false);
completeBlock(lastBlock, iip, false);
}
return committed;
}
@ -841,11 +843,15 @@ public class BlockManager implements BlockStatsMXBean {
/**
* Convert a specified block of the file to a complete block.
* @param curBlock - block to be completed
* @param iip - INodes in path to file containing curBlock; if null,
* this will be resolved internally
* @param force - force completion of the block
* @throws IOException if the block does not have at least a minimal number
* of replicas reported from data-nodes.
*/
private void completeBlock(BlockInfo curBlock, boolean force)
throws IOException {
private void completeBlock(BlockInfo curBlock, INodesInPath iip,
boolean force) throws IOException {
if (curBlock.isComplete()) {
return;
}
@ -860,7 +866,8 @@ public class BlockManager implements BlockStatsMXBean {
"Cannot complete block: block has not been COMMITTED by the client");
}
curBlock.convertToCompleteBlock();
convertToCompleteBlock(curBlock, iip);
// Since safe-mode only counts complete blocks, and we now have
// one more complete block, we need to adjust the total up, and
// also count it as safe, if we have at least the minimum replica
@ -874,6 +881,22 @@ public class BlockManager implements BlockStatsMXBean {
curBlock);
}
/**
* Convert a specified block of the file to a complete block.
* Skips validity checking and safe mode block total updates; use
* {@link BlockManager#completeBlock} to include these.
* @param curBlock - block to be completed
* @param iip - INodes in path to file containing curBlock; if null,
* this will be resolved internally
* @throws IOException if the block does not have at least a minimal number
* of replicas reported from data-nodes.
*/
private void convertToCompleteBlock(BlockInfo curBlock, INodesInPath iip)
throws IOException {
curBlock.convertToCompleteBlock();
namesystem.getFSDirectory().updateSpaceForCompleteBlock(curBlock, iip);
}
/**
* Force the given block in the given file to be marked as complete,
* regardless of whether enough replicas are present. This is necessary
@ -881,7 +904,7 @@ public class BlockManager implements BlockStatsMXBean {
*/
public void forceCompleteBlock(final BlockInfo block) throws IOException {
block.commitBlock(block);
completeBlock(block, true);
completeBlock(block, null, true);
}
/**
@ -2910,7 +2933,7 @@ public class BlockManager implements BlockStatsMXBean {
int numCurrentReplica = countLiveNodes(storedBlock);
if (storedBlock.getBlockUCState() == BlockUCState.COMMITTED
&& hasMinStorage(storedBlock, numCurrentReplica)) {
completeBlock(storedBlock, false);
completeBlock(storedBlock, null, false);
} else if (storedBlock.isComplete() && result == AddBlockResult.ADDED) {
// check whether safe replication is reached for the block
// only complete blocks are counted towards that.
@ -2985,7 +3008,7 @@ public class BlockManager implements BlockStatsMXBean {
if(storedBlock.getBlockUCState() == BlockUCState.COMMITTED &&
hasMinStorage(storedBlock, numLiveReplicas)) {
addExpectedReplicasToPending(storedBlock);
completeBlock(storedBlock, false);
completeBlock(storedBlock, null, false);
} else if (storedBlock.isComplete() && result == AddBlockResult.ADDED) {
// check whether safe replication is reached for the block
// only complete blocks are counted towards that

View File

@ -418,9 +418,9 @@ public class FSDirAttrOp {
// Make sure the directory has sufficient quotas
short oldBR = file.getPreferredBlockReplication();
long size = file.computeFileSize(true, true);
// Ensure the quota does not exceed
if (oldBR < replication) {
long size = file.computeFileSize(true, true);
fsd.updateCount(iip, 0L, size, oldBR, replication, true);
}
@ -428,14 +428,10 @@ public class FSDirAttrOp {
short targetReplication = (short) Math.max(
replication, file.getPreferredBlockReplication());
if (oldBR > replication) {
fsd.updateCount(iip, 0L, size, oldBR, targetReplication, true);
}
for (BlockInfo b : file.getBlocks()) {
if (oldBR == targetReplication) {
continue;
}
if (oldBR > replication) {
fsd.updateCount(iip, 0L, b.getNumBytes(), oldBR, targetReplication,
true);
}
bm.setReplication(oldBR, targetReplication, b);
}

View File

@ -38,6 +38,7 @@ import org.apache.hadoop.fs.permission.FsPermission;
import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.hdfs.DFSUtil;
import org.apache.hadoop.hdfs.protocol.BlockStoragePolicy;
import org.apache.hadoop.hdfs.protocol.ErasureCodingPolicy;
import org.apache.hadoop.hdfs.protocol.FSLimitException.MaxDirectoryItemsExceededException;
import org.apache.hadoop.hdfs.protocol.FSLimitException.PathComponentTooLongException;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
@ -47,6 +48,7 @@ import org.apache.hadoop.hdfs.protocol.SnapshotAccessControlException;
import org.apache.hadoop.hdfs.protocol.proto.HdfsProtos;
import org.apache.hadoop.hdfs.protocolPB.PBHelperClient;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoStriped;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockStoragePolicySuite;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
@ -930,6 +932,51 @@ public class FSDirectory implements Closeable {
}
}
/**
* Update the cached quota space for a block that is being completed.
* Must only be called once, as the block is being completed.
* @param completeBlk - Completed block for which to update space
* @param inodes - INodes in path to file containing completeBlk; if null
* this will be resolved internally
*/
public void updateSpaceForCompleteBlock(BlockInfo completeBlk,
INodesInPath inodes) throws IOException {
assert namesystem.hasWriteLock();
INodesInPath iip = inodes != null ? inodes :
INodesInPath.fromINode(namesystem.getBlockCollection(completeBlk));
INodeFile fileINode = iip.getLastINode().asFile();
// Adjust disk space consumption if required
final long diff;
final short replicationFactor;
if (fileINode.isStriped()) {
final ErasureCodingPolicy ecPolicy =
FSDirErasureCodingOp.getErasureCodingPolicy(namesystem, iip);
final short numDataUnits = (short) ecPolicy.getNumDataUnits();
final short numParityUnits = (short) ecPolicy.getNumParityUnits();
final long numBlocks = numDataUnits + numParityUnits;
final long fullBlockGroupSize =
fileINode.getPreferredBlockSize() * numBlocks;
final BlockInfoStriped striped =
new BlockInfoStriped(completeBlk, ecPolicy);
final long actualBlockGroupSize = striped.spaceConsumed();
diff = fullBlockGroupSize - actualBlockGroupSize;
replicationFactor = (short) 1;
} else {
diff = fileINode.getPreferredBlockSize() - completeBlk.getNumBytes();
replicationFactor = fileINode.getFileReplication();
}
if (diff > 0) {
try {
updateSpaceConsumed(iip, 0, -diff, replicationFactor);
} catch (IOException e) {
LOG.warn("Unexpected exception while updating disk space.", e);
}
}
}
public EnumCounters<StorageType> getStorageTypeDeltas(byte storagePolicyID,
long dsDelta, short oldRep, short newRep) {
EnumCounters<StorageType> typeSpaceDeltas =

View File

@ -186,7 +186,6 @@ import org.apache.hadoop.hdfs.protocol.DirectoryListing;
import org.apache.hadoop.hdfs.protocol.ErasureCodingPolicy;
import org.apache.hadoop.hdfs.protocol.EncryptionZone;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
import org.apache.hadoop.hdfs.protocol.HdfsConstants.DatanodeReportType;
import org.apache.hadoop.hdfs.protocol.HdfsConstants.SafeModeAction;
import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
@ -207,7 +206,6 @@ import org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenSecretMan
import org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenSecretManager.SecretManagerState;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockCollection;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoStriped;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockUnderConstructionFeature;
import org.apache.hadoop.hdfs.server.blockmanagement.DatanodeDescriptor;
@ -3292,40 +3290,7 @@ public class FSNamesystem implements Namesystem, FSNamesystemMBean,
final Block commitBlock) throws IOException {
assert hasWriteLock();
Preconditions.checkArgument(fileINode.isUnderConstruction());
if (!blockManager.commitOrCompleteLastBlock(fileINode, commitBlock)) {
return;
}
// Adjust disk space consumption if required
final long diff;
final short replicationFactor;
if (fileINode.isStriped()) {
final ErasureCodingPolicy ecPolicy = FSDirErasureCodingOp
.getErasureCodingPolicy(this, iip);
final short numDataUnits = (short) ecPolicy.getNumDataUnits();
final short numParityUnits = (short) ecPolicy.getNumParityUnits();
final long numBlocks = numDataUnits + numParityUnits;
final long fullBlockGroupSize =
fileINode.getPreferredBlockSize() * numBlocks;
final BlockInfoStriped striped = new BlockInfoStriped(commitBlock,
ecPolicy);
final long actualBlockGroupSize = striped.spaceConsumed();
diff = fullBlockGroupSize - actualBlockGroupSize;
replicationFactor = (short) 1;
} else {
diff = fileINode.getPreferredBlockSize() - commitBlock.getNumBytes();
replicationFactor = fileINode.getFileReplication();
}
if (diff > 0) {
try {
dir.updateSpaceConsumed(iip, 0, -diff, replicationFactor);
} catch (IOException e) {
LOG.warn("Unexpected exception while updating disk space.", e);
}
}
blockManager.commitOrCompleteLastBlock(fileINode, commitBlock, iip);
}
void addCommittedBlocksToPending(final INodeFile pendingFile) {
@ -3583,7 +3548,6 @@ public class FSNamesystem implements Namesystem, FSNamesystemMBean,
/**
* @param pendingFile open file that needs to be closed
* @param storedBlock last block
* @return Path of the file that was closed.
* @throws IOException on error
*/
@VisibleForTesting
@ -5734,6 +5698,7 @@ public class FSNamesystem implements Namesystem, FSNamesystemMBean,
}
/** @return the FSDirectory. */
@Override
public FSDirectory getFSDirectory() {
return dir;
}

View File

@ -30,6 +30,8 @@ public interface Namesystem extends RwLock, SafeMode {
BlockCollection getBlockCollection(long id);
FSDirectory getFSDirectory();
void startSecretManagerIfNecessary();
boolean isInSnapshot(long blockCollectionID);

View File

@ -18,10 +18,14 @@
package org.apache.hadoop.hdfs.server.namenode;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertTrue;
import java.io.IOException;
import java.util.ArrayList;
import java.util.EnumSet;
import java.util.HashMap;
import java.util.List;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.ContentSummary;
@ -33,61 +37,79 @@ import org.apache.hadoop.hdfs.DFSOutputStream;
import org.apache.hadoop.hdfs.DFSTestUtil;
import org.apache.hadoop.hdfs.DistributedFileSystem;
import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.TestFileCreation;
import org.apache.hadoop.hdfs.client.HdfsDataOutputStream;
import org.apache.hadoop.hdfs.protocol.DSQuotaExceededException;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
import org.apache.hadoop.hdfs.protocol.QuotaByStorageTypeExceededException;
import org.apache.hadoop.hdfs.server.datanode.InternalDataNodeTestUtils;
import org.apache.hadoop.hdfs.server.namenode.snapshot.Snapshot;
import org.apache.hadoop.ipc.RemoteException;
import org.junit.After;
import org.apache.hadoop.hdfs.protocolPB.DatanodeProtocolClientSideTranslatorPB;
import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.StorageReceivedDeletedBlocks;
import org.apache.hadoop.test.GenericTestUtils;
import org.apache.hadoop.test.GenericTestUtils.LogCapturer;
import org.junit.AfterClass;
import org.junit.Assert;
import org.junit.Before;
import org.junit.BeforeClass;
import org.junit.Test;
import org.mockito.Mockito;
import org.mockito.invocation.InvocationOnMock;
import org.mockito.stubbing.Answer;
public class TestDiskspaceQuotaUpdate {
private static final int BLOCKSIZE = 1024;
private static final short REPLICATION = 4;
static final long seed = 0L;
private static final Path dir = new Path("/TestQuotaUpdate");
private static final Path BASE_DIR = new Path("/TestQuotaUpdate");
private Configuration conf;
private MiniDFSCluster cluster;
private FSDirectory fsdir;
private DistributedFileSystem dfs;
private static Configuration conf;
private static MiniDFSCluster cluster;
@Before
public void setUp() throws Exception {
@BeforeClass
public static void setUp() throws Exception {
conf = new Configuration();
conf.setLong(DFSConfigKeys.DFS_BLOCK_SIZE_KEY, BLOCKSIZE);
cluster = new MiniDFSCluster.Builder(conf).numDataNodes(REPLICATION)
.build();
cluster.waitActive();
fsdir = cluster.getNamesystem().getFSDirectory();
dfs = cluster.getFileSystem();
}
@After
public void tearDown() throws Exception {
@AfterClass
public static void tearDown() throws Exception {
if (cluster != null) {
cluster.shutdown();
cluster = null;
}
}
private Path getParent(String testName) {
return new Path(BASE_DIR, testName);
}
private FSDirectory getFSDirectory() {
return cluster.getNamesystem().getFSDirectory();
}
private DistributedFileSystem getDFS() throws IOException {
return cluster.getFileSystem();
}
/**
* Test if the quota can be correctly updated for create file
*/
@Test (timeout=60000)
public void testQuotaUpdateWithFileCreate() throws Exception {
final Path foo = new Path(dir, "foo");
final Path foo =
new Path(getParent(GenericTestUtils.getMethodName()), "foo");
Path createdFile = new Path(foo, "created_file.data");
dfs.mkdirs(foo);
dfs.setQuota(foo, Long.MAX_VALUE-1, Long.MAX_VALUE-1);
getDFS().mkdirs(foo);
getDFS().setQuota(foo, Long.MAX_VALUE-1, Long.MAX_VALUE-1);
long fileLen = BLOCKSIZE * 2 + BLOCKSIZE / 2;
DFSTestUtil.createFile(dfs, createdFile, BLOCKSIZE / 16,
DFSTestUtil.createFile(getDFS(), createdFile, BLOCKSIZE / 16,
fileLen, BLOCKSIZE, REPLICATION, seed);
INode fnode = fsdir.getINode4Write(foo.toString());
INode fnode = getFSDirectory().getINode4Write(foo.toString());
assertTrue(fnode.isDirectory());
assertTrue(fnode.isQuotaSet());
QuotaCounts cnt = fnode.asDirectory().getDirectoryWithQuotaFeature()
@ -101,18 +123,20 @@ public class TestDiskspaceQuotaUpdate {
*/
@Test (timeout=60000)
public void testUpdateQuotaForAppend() throws Exception {
final Path foo = new Path(dir ,"foo");
final Path foo =
new Path(getParent(GenericTestUtils.getMethodName()), "foo");
final Path bar = new Path(foo, "bar");
long currentFileLen = BLOCKSIZE;
DFSTestUtil.createFile(dfs, bar, currentFileLen, REPLICATION, seed);
dfs.setQuota(foo, Long.MAX_VALUE-1, Long.MAX_VALUE-1);
DFSTestUtil.createFile(getDFS(), bar, currentFileLen, REPLICATION, seed);
getDFS().setQuota(foo, Long.MAX_VALUE-1, Long.MAX_VALUE-1);
// append half of the block data, the previous file length is at block
// boundary
DFSTestUtil.appendFile(dfs, bar, BLOCKSIZE / 2);
DFSTestUtil.appendFile(getDFS(), bar, BLOCKSIZE / 2);
currentFileLen += (BLOCKSIZE / 2);
INodeDirectory fooNode = fsdir.getINode4Write(foo.toString()).asDirectory();
INodeDirectory fooNode =
getFSDirectory().getINode4Write(foo.toString()).asDirectory();
assertTrue(fooNode.isQuotaSet());
QuotaCounts quota = fooNode.getDirectoryWithQuotaFeature()
.getSpaceConsumed();
@ -120,11 +144,11 @@ public class TestDiskspaceQuotaUpdate {
long ds = quota.getStorageSpace();
assertEquals(2, ns); // foo and bar
assertEquals(currentFileLen * REPLICATION, ds);
ContentSummary c = dfs.getContentSummary(foo);
ContentSummary c = getDFS().getContentSummary(foo);
assertEquals(c.getSpaceConsumed(), ds);
// append another block, the previous file length is not at block boundary
DFSTestUtil.appendFile(dfs, bar, BLOCKSIZE);
DFSTestUtil.appendFile(getDFS(), bar, BLOCKSIZE);
currentFileLen += BLOCKSIZE;
quota = fooNode.getDirectoryWithQuotaFeature().getSpaceConsumed();
@ -132,11 +156,11 @@ public class TestDiskspaceQuotaUpdate {
ds = quota.getStorageSpace();
assertEquals(2, ns); // foo and bar
assertEquals(currentFileLen * REPLICATION, ds);
c = dfs.getContentSummary(foo);
c = getDFS().getContentSummary(foo);
assertEquals(c.getSpaceConsumed(), ds);
// append several blocks
DFSTestUtil.appendFile(dfs, bar, BLOCKSIZE * 3 + BLOCKSIZE / 8);
DFSTestUtil.appendFile(getDFS(), bar, BLOCKSIZE * 3 + BLOCKSIZE / 8);
currentFileLen += (BLOCKSIZE * 3 + BLOCKSIZE / 8);
quota = fooNode.getDirectoryWithQuotaFeature().getSpaceConsumed();
@ -144,7 +168,7 @@ public class TestDiskspaceQuotaUpdate {
ds = quota.getStorageSpace();
assertEquals(2, ns); // foo and bar
assertEquals(currentFileLen * REPLICATION, ds);
c = dfs.getContentSummary(foo);
c = getDFS().getContentSummary(foo);
assertEquals(c.getSpaceConsumed(), ds);
}
@ -154,16 +178,18 @@ public class TestDiskspaceQuotaUpdate {
*/
@Test (timeout=60000)
public void testUpdateQuotaForFSync() throws Exception {
final Path foo = new Path("/foo");
final Path foo =
new Path(getParent(GenericTestUtils.getMethodName()), "foo");
final Path bar = new Path(foo, "bar");
DFSTestUtil.createFile(dfs, bar, BLOCKSIZE, REPLICATION, 0L);
dfs.setQuota(foo, Long.MAX_VALUE - 1, Long.MAX_VALUE - 1);
DFSTestUtil.createFile(getDFS(), bar, BLOCKSIZE, REPLICATION, 0L);
getDFS().setQuota(foo, Long.MAX_VALUE - 1, Long.MAX_VALUE - 1);
FSDataOutputStream out = dfs.append(bar);
FSDataOutputStream out = getDFS().append(bar);
out.write(new byte[BLOCKSIZE / 4]);
((DFSOutputStream) out.getWrappedStream()).hsync(EnumSet.of(HdfsDataOutputStream.SyncFlag.UPDATE_LENGTH));
INodeDirectory fooNode = fsdir.getINode4Write(foo.toString()).asDirectory();
INodeDirectory fooNode =
getFSDirectory().getINode4Write(foo.toString()).asDirectory();
QuotaCounts quota = fooNode.getDirectoryWithQuotaFeature()
.getSpaceConsumed();
long ns = quota.getNameSpace();
@ -174,7 +200,7 @@ public class TestDiskspaceQuotaUpdate {
out.write(new byte[BLOCKSIZE / 4]);
out.close();
fooNode = fsdir.getINode4Write(foo.toString()).asDirectory();
fooNode = getFSDirectory().getINode4Write(foo.toString()).asDirectory();
quota = fooNode.getDirectoryWithQuotaFeature().getSpaceConsumed();
ns = quota.getNameSpace();
ds = quota.getStorageSpace();
@ -182,7 +208,7 @@ public class TestDiskspaceQuotaUpdate {
assertEquals((BLOCKSIZE + BLOCKSIZE / 2) * REPLICATION, ds);
// append another block
DFSTestUtil.appendFile(dfs, bar, BLOCKSIZE);
DFSTestUtil.appendFile(getDFS(), bar, BLOCKSIZE);
quota = fooNode.getDirectoryWithQuotaFeature().getSpaceConsumed();
ns = quota.getNameSpace();
@ -196,21 +222,21 @@ public class TestDiskspaceQuotaUpdate {
*/
@Test (timeout=60000)
public void testAppendOverStorageQuota() throws Exception {
final Path dir = new Path("/TestAppendOverQuota");
final Path dir = getParent(GenericTestUtils.getMethodName());
final Path file = new Path(dir, "file");
// create partial block file
dfs.mkdirs(dir);
DFSTestUtil.createFile(dfs, file, BLOCKSIZE/2, REPLICATION, seed);
getDFS().mkdirs(dir);
DFSTestUtil.createFile(getDFS(), file, BLOCKSIZE/2, REPLICATION, seed);
// lower quota to cause exception when appending to partial block
dfs.setQuota(dir, Long.MAX_VALUE - 1, 1);
final INodeDirectory dirNode = fsdir.getINode4Write(dir.toString())
.asDirectory();
getDFS().setQuota(dir, Long.MAX_VALUE - 1, 1);
final INodeDirectory dirNode =
getFSDirectory().getINode4Write(dir.toString()).asDirectory();
final long spaceUsed = dirNode.getDirectoryWithQuotaFeature()
.getSpaceConsumed().getStorageSpace();
try {
DFSTestUtil.appendFile(dfs, file, BLOCKSIZE);
DFSTestUtil.appendFile(getDFS(), file, BLOCKSIZE);
Assert.fail("append didn't fail");
} catch (DSQuotaExceededException e) {
// ignore
@ -218,7 +244,7 @@ public class TestDiskspaceQuotaUpdate {
LeaseManager lm = cluster.getNamesystem().getLeaseManager();
// check that the file exists, isn't UC, and has no dangling lease
INodeFile inode = fsdir.getINode(file.toString()).asFile();
INodeFile inode = getFSDirectory().getINode(file.toString()).asFile();
Assert.assertNotNull(inode);
Assert.assertFalse("should not be UC", inode.isUnderConstruction());
Assert.assertNull("should not have a lease", lm.getLease(inode));
@ -227,7 +253,7 @@ public class TestDiskspaceQuotaUpdate {
.getSpaceConsumed().getStorageSpace();
assertEquals(spaceUsed, newSpaceUsed);
// make sure edits aren't corrupted
dfs.recoverLease(file);
getDFS().recoverLease(file);
cluster.restartNameNodes();
}
@ -237,23 +263,23 @@ public class TestDiskspaceQuotaUpdate {
*/
@Test (timeout=60000)
public void testAppendOverTypeQuota() throws Exception {
final Path dir = new Path("/TestAppendOverTypeQuota");
final Path dir = getParent(GenericTestUtils.getMethodName());
final Path file = new Path(dir, "file");
// create partial block file
dfs.mkdirs(dir);
getDFS().mkdirs(dir);
// set the storage policy on dir
dfs.setStoragePolicy(dir, HdfsConstants.ONESSD_STORAGE_POLICY_NAME);
DFSTestUtil.createFile(dfs, file, BLOCKSIZE/2, REPLICATION, seed);
getDFS().setStoragePolicy(dir, HdfsConstants.ONESSD_STORAGE_POLICY_NAME);
DFSTestUtil.createFile(getDFS(), file, BLOCKSIZE/2, REPLICATION, seed);
// set quota of SSD to 1L
dfs.setQuotaByStorageType(dir, StorageType.SSD, 1L);
final INodeDirectory dirNode = fsdir.getINode4Write(dir.toString())
.asDirectory();
getDFS().setQuotaByStorageType(dir, StorageType.SSD, 1L);
final INodeDirectory dirNode =
getFSDirectory().getINode4Write(dir.toString()).asDirectory();
final long spaceUsed = dirNode.getDirectoryWithQuotaFeature()
.getSpaceConsumed().getStorageSpace();
try {
DFSTestUtil.appendFile(dfs, file, BLOCKSIZE);
DFSTestUtil.appendFile(getDFS(), file, BLOCKSIZE);
Assert.fail("append didn't fail");
} catch (QuotaByStorageTypeExceededException e) {
//ignore
@ -261,7 +287,7 @@ public class TestDiskspaceQuotaUpdate {
// check that the file exists, isn't UC, and has no dangling lease
LeaseManager lm = cluster.getNamesystem().getLeaseManager();
INodeFile inode = fsdir.getINode(file.toString()).asFile();
INodeFile inode = getFSDirectory().getINode(file.toString()).asFile();
Assert.assertNotNull(inode);
Assert.assertFalse("should not be UC", inode.isUnderConstruction());
Assert.assertNull("should not have a lease", lm.getLease(inode));
@ -270,7 +296,7 @@ public class TestDiskspaceQuotaUpdate {
.getSpaceConsumed().getStorageSpace();
assertEquals(spaceUsed, newSpaceUsed);
// make sure edits aren't corrupted
dfs.recoverLease(file);
getDFS().recoverLease(file);
cluster.restartNameNodes();
}
@ -279,21 +305,21 @@ public class TestDiskspaceQuotaUpdate {
*/
@Test (timeout=60000)
public void testTruncateOverQuota() throws Exception {
final Path dir = new Path("/TestTruncateOverquota");
final Path dir = getParent(GenericTestUtils.getMethodName());
final Path file = new Path(dir, "file");
// create partial block file
dfs.mkdirs(dir);
DFSTestUtil.createFile(dfs, file, BLOCKSIZE/2, REPLICATION, seed);
getDFS().mkdirs(dir);
DFSTestUtil.createFile(getDFS(), file, BLOCKSIZE/2, REPLICATION, seed);
// lower quota to cause exception when appending to partial block
dfs.setQuota(dir, Long.MAX_VALUE - 1, 1);
final INodeDirectory dirNode = fsdir.getINode4Write(dir.toString())
.asDirectory();
getDFS().setQuota(dir, Long.MAX_VALUE - 1, 1);
final INodeDirectory dirNode =
getFSDirectory().getINode4Write(dir.toString()).asDirectory();
final long spaceUsed = dirNode.getDirectoryWithQuotaFeature()
.getSpaceConsumed().getStorageSpace();
try {
dfs.truncate(file, BLOCKSIZE / 2 - 1);
getDFS().truncate(file, BLOCKSIZE / 2 - 1);
Assert.fail("truncate didn't fail");
} catch (RemoteException e) {
assertTrue(e.getClassName().contains("DSQuotaExceededException"));
@ -301,7 +327,7 @@ public class TestDiskspaceQuotaUpdate {
// check that the file exists, isn't UC, and has no dangling lease
LeaseManager lm = cluster.getNamesystem().getLeaseManager();
INodeFile inode = fsdir.getINode(file.toString()).asFile();
INodeFile inode = getFSDirectory().getINode(file.toString()).asFile();
Assert.assertNotNull(inode);
Assert.assertFalse("should not be UC", inode.isUnderConstruction());
Assert.assertNull("should not have a lease", lm.getLease(inode));
@ -310,7 +336,7 @@ public class TestDiskspaceQuotaUpdate {
.getSpaceConsumed().getStorageSpace();
assertEquals(spaceUsed, newSpaceUsed);
// make sure edits aren't corrupted
dfs.recoverLease(file);
getDFS().recoverLease(file);
cluster.restartNameNodes();
}
@ -320,33 +346,34 @@ public class TestDiskspaceQuotaUpdate {
@Test
public void testQuotaInitialization() throws Exception {
final int size = 500;
Path testDir = new Path("/testDir");
Path testDir =
new Path(getParent(GenericTestUtils.getMethodName()), "testDir");
long expectedSize = 3 * BLOCKSIZE + BLOCKSIZE/2;
dfs.mkdirs(testDir);
dfs.setQuota(testDir, size*4, expectedSize*size*2);
getDFS().mkdirs(testDir);
getDFS().setQuota(testDir, size*4, expectedSize*size*2);
Path[] testDirs = new Path[size];
for (int i = 0; i < size; i++) {
testDirs[i] = new Path(testDir, "sub" + i);
dfs.mkdirs(testDirs[i]);
dfs.setQuota(testDirs[i], 100, 1000000);
DFSTestUtil.createFile(dfs, new Path(testDirs[i], "a"), expectedSize,
getDFS().mkdirs(testDirs[i]);
getDFS().setQuota(testDirs[i], 100, 1000000);
DFSTestUtil.createFile(getDFS(), new Path(testDirs[i], "a"), expectedSize,
(short)1, 1L);
}
// Directly access the name system to obtain the current cached usage.
INodeDirectory root = fsdir.getRoot();
INodeDirectory root = getFSDirectory().getRoot();
HashMap<String, Long> nsMap = new HashMap<String, Long>();
HashMap<String, Long> dsMap = new HashMap<String, Long>();
scanDirsWithQuota(root, nsMap, dsMap, false);
fsdir.updateCountForQuota(1);
getFSDirectory().updateCountForQuota(1);
scanDirsWithQuota(root, nsMap, dsMap, true);
fsdir.updateCountForQuota(2);
getFSDirectory().updateCountForQuota(2);
scanDirsWithQuota(root, nsMap, dsMap, true);
fsdir.updateCountForQuota(4);
getFSDirectory().updateCountForQuota(4);
scanDirsWithQuota(root, nsMap, dsMap, true);
}
@ -372,4 +399,130 @@ public class TestDiskspaceQuotaUpdate {
}
}
}
/**
* Test that the cached quota stays correct between the COMMIT
* and COMPLETE block steps, even if the replication factor is
* changed during this time.
*/
@Test (timeout=60000)
public void testQuotaIssuesWhileCommitting() throws Exception {
// We want a one-DN cluster so that we can force a lack of
// commit by only instrumenting a single DN; we kill the other 3
List<MiniDFSCluster.DataNodeProperties> dnprops = new ArrayList<>();
try {
for (int i = REPLICATION - 1; i > 0; i--) {
dnprops.add(cluster.stopDataNode(i));
}
DatanodeProtocolClientSideTranslatorPB nnSpy =
InternalDataNodeTestUtils.spyOnBposToNN(
cluster.getDataNodes().get(0), cluster.getNameNode());
testQuotaIssuesWhileCommittingHelper(nnSpy, (short) 1, (short) 4);
testQuotaIssuesWhileCommittingHelper(nnSpy, (short) 4, (short) 1);
// Don't actually change replication; just check that the sizes
// agree during the commit period
testQuotaIssuesWhileCommittingHelper(nnSpy, (short) 1, (short) 1);
} finally {
for (MiniDFSCluster.DataNodeProperties dnprop : dnprops) {
cluster.restartDataNode(dnprop, true);
}
cluster.waitActive();
}
}
private void testQuotaIssuesWhileCommittingHelper(
DatanodeProtocolClientSideTranslatorPB nnSpy,
final short initialReplication, final short finalReplication)
throws Exception {
final String logStmt =
"BUG: Inconsistent storagespace for directory";
final Path dir = new Path(getParent(GenericTestUtils.getMethodName()),
String.format("%d-%d", initialReplication, finalReplication));
final Path file = new Path(dir, "testfile");
LogCapturer logs = LogCapturer.captureLogs(NameNode.LOG);
Mockito.doAnswer(new Answer<Object>() {
@Override
public Object answer(InvocationOnMock invocation) throws Throwable {
if (finalReplication != initialReplication) {
getDFS().setReplication(file, finalReplication);
}
// Call getContentSummary before the DN can notify the NN
// that the block has been received to check for discrepancy
getDFS().getContentSummary(dir);
invocation.callRealMethod();
return null;
}
}).when(nnSpy).blockReceivedAndDeleted(
Mockito.<DatanodeRegistration>anyObject(),
Mockito.anyString(),
Mockito.<StorageReceivedDeletedBlocks[]>anyObject()
);
getDFS().mkdirs(dir);
getDFS().setQuota(dir, Long.MAX_VALUE - 1, Long.MAX_VALUE - 1);
DFSTestUtil.createFile(getDFS(), file, BLOCKSIZE/2, initialReplication, 1L);
// Also check for discrepancy after completing the file
getDFS().getContentSummary(dir);
assertFalse(logs.getOutput().contains(logStmt));
}
/**
* Test that the cached quota remains correct when the block has been
* written to but not yet committed, even if the replication factor
* is updated during this time.
*/
private void testQuotaIssuesBeforeCommitting(short initialReplication,
short finalReplication) throws Exception {
final String logStmt =
"BUG: Inconsistent storagespace for directory";
final Path dir = new Path(getParent(GenericTestUtils.getMethodName()),
String.format("%d-%d", initialReplication, finalReplication));
final Path file = new Path(dir, "testfile");
LogCapturer logs = LogCapturer.captureLogs(NameNode.LOG);
getDFS().mkdirs(dir);
getDFS().setQuota(dir, Long.MAX_VALUE - 1, Long.MAX_VALUE - 1);
FSDataOutputStream out =
TestFileCreation.createFile(getDFS(), file, initialReplication);
TestFileCreation.writeFile(out, BLOCKSIZE / 2);
out.hflush();
getDFS().getContentSummary(dir);
if (finalReplication != initialReplication) {
// While the block is visible to the NN but has not yet been committed,
// change the replication
getDFS().setReplication(file, finalReplication);
}
out.close();
getDFS().getContentSummary(dir);
assertFalse(logs.getOutput().contains(logStmt));
}
@Test (timeout=60000)
public void testCachedComputedSizesAgreeBeforeCommitting() throws Exception {
// Don't actually change replication; just check that the sizes
// agree before the commit period
testQuotaIssuesBeforeCommitting((short)1, (short)1);
}
@Test (timeout=60000)
public void testDecreaseReplicationBeforeCommitting() throws Exception {
testQuotaIssuesBeforeCommitting((short)4, (short)1);
}
@Test (timeout=60000)
public void testIncreaseReplicationBeforeCommitting() throws Exception {
testQuotaIssuesBeforeCommitting((short)1, (short)4);
}
}