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

Revert "HBASE-18814 Improve TableSnapshotInputFormat to allow more multiple mappers per region" due to wrong jira id. 

This reverts commit f20580a530.
This commit is contained in:
Ashu Pachauri 2017-09-30 01:43:10 -07:00
parent cacf3f5835
commit 367dfabf06
12 changed files with 50 additions and 381 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
hbase-it/src/test/java/org/apache/hadoop/hbase/mapreduce/IntegrationTestTableSnapshotInputFormat.java
View File

@ -151,7 +151,7 @@ public class IntegrationTestTableSnapshotInputFormat extends IntegrationTestBase
int expectedNumSplits = numRegions > 2 ? numRegions - 2 : numRegions;
org.apache.hadoop.hbase.mapreduce.TestTableSnapshotInputFormat.doTestWithMapReduce(util,
tableName, snapshotName, START_ROW, END_ROW, tableDir, numRegions, 1,
tableName, snapshotName, START_ROW, END_ROW, tableDir, numRegions,
expectedNumSplits, false);
} else if (mr.equalsIgnoreCase(MAPRED_IMPLEMENTATION)) {
/*
@ -165,7 +165,7 @@ public class IntegrationTestTableSnapshotInputFormat extends IntegrationTestBase
int expectedNumSplits = numRegions;
org.apache.hadoop.hbase.mapred.TestTableSnapshotInputFormat.doTestWithMapReduce(util,
tableName, snapshotName, MAPRED_START_ROW, MAPRED_END_ROW, tableDir, numRegions, 1,
tableName, snapshotName, MAPRED_START_ROW, MAPRED_END_ROW, tableDir, numRegions,
expectedNumSplits, false);
} else {
throw new IllegalArgumentException("Unrecognized mapreduce implementation: " + mr +".");

38
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapred/TableMapReduceUtil.java
View File

@ -33,7 +33,6 @@ import org.apache.hadoop.hbase.mapreduce.ResultSerialization;
import org.apache.hadoop.hbase.security.User;
import org.apache.hadoop.hbase.security.UserProvider;
import org.apache.hadoop.hbase.security.token.TokenUtil;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.mapred.FileInputFormat;
import org.apache.hadoop.mapred.InputFormat;
import org.apache.hadoop.mapred.JobConf;
@ -186,43 +185,6 @@ public class TableMapReduceUtil {
org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.resetCacheConfig(job);
}
/**
* Sets up the job for reading from a table snapshot. It bypasses hbase servers
* and read directly from snapshot files.
*
* @param snapshotName The name of the snapshot (of a table) to read from.
* @param columns The columns to scan.
* @param mapper The mapper class to use.
* @param outputKeyClass The class of the output key.
* @param outputValueClass The class of the output value.
* @param jobConf The current job to adjust. Make sure the passed job is
* carrying all necessary HBase configuration.
* @param addDependencyJars upload HBase jars and jars for any of the configured
* job classes via the distributed cache (tmpjars).
* @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should
* have write permissions to this directory, and this should not be a subdirectory of rootdir.
* After the job is finished, restore directory can be deleted.
* @param splitAlgo algorithm to split
* @param numSplitsPerRegion how many input splits to generate per one region
* @throws IOException When setting up the details fails.
* @see TableSnapshotInputFormat
*/
public static void initTableSnapshotMapJob(String snapshotName, String columns,
Class<? extends TableMap> mapper,
Class<?> outputKeyClass,
Class<?> outputValueClass, JobConf jobConf,
boolean addDependencyJars, Path tmpRestoreDir,
RegionSplitter.SplitAlgorithm splitAlgo,
int numSplitsPerRegion)
throws IOException {
TableSnapshotInputFormat.setInput(jobConf, snapshotName, tmpRestoreDir, splitAlgo,
numSplitsPerRegion);
initTableMapJob(snapshotName, columns, mapper, outputKeyClass, outputValueClass, jobConf,
addDependencyJars, TableSnapshotInputFormat.class);
org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.resetCacheConfig(jobConf);
}
/**
* Use this before submitting a TableReduce job. It will
* appropriately set up the JobConf.

18
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapred/TableSnapshotInputFormat.java
View File

@ -26,13 +26,11 @@ import org.apache.hadoop.hbase.client.Result;
import org.apache.hadoop.hbase.client.Scan;
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
import org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormatImpl;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.mapred.InputFormat;
import org.apache.hadoop.mapred.InputSplit;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.RecordReader;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.mapreduce.Job;
import java.io.DataInput;
import java.io.DataOutput;
@ -165,20 +163,4 @@ public class TableSnapshotInputFormat implements InputFormat<ImmutableBytesWrita
throws IOException {
TableSnapshotInputFormatImpl.setInput(job, snapshotName, restoreDir);
}
/**
* Configures the job to use TableSnapshotInputFormat to read from a snapshot.
* @param job the job to configure
* @param snapshotName the name of the snapshot to read from
* @param restoreDir a temporary directory to restore the snapshot into. Current user should
* have write permissions to this directory, and this should not be a subdirectory of rootdir.
* After the job is finished, restoreDir can be deleted.
* @param splitAlgo split algorithm to generate splits from region
* @param numSplitsPerRegion how many input splits to generate per one region
* @throws IOException if an error occurs
*/
public static void setInput(JobConf job, String snapshotName, Path restoreDir,
RegionSplitter.SplitAlgorithm splitAlgo, int numSplitsPerRegion) throws IOException {
TableSnapshotInputFormatImpl.setInput(job, snapshotName, restoreDir, splitAlgo, numSplitsPerRegion);
}
}

38
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java
View File

@ -55,7 +55,6 @@ import org.apache.hadoop.hbase.security.UserProvider;
import org.apache.hadoop.hbase.security.token.TokenUtil;
import org.apache.hadoop.hbase.util.Base64;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.hbase.zookeeper.ZKConfig;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapreduce.InputFormat;
@ -376,43 +375,6 @@ public class TableMapReduceUtil {
resetCacheConfig(job.getConfiguration());
}
/**
* Sets up the job for reading from a table snapshot. It bypasses hbase servers
* and read directly from snapshot files.
*
* @param snapshotName The name of the snapshot (of a table) to read from.
* @param scan The scan instance with the columns, time range etc.
* @param mapper The mapper class to use.
* @param outputKeyClass The class of the output key.
* @param outputValueClass The class of the output value.
* @param job The current job to adjust. Make sure the passed job is
* carrying all necessary HBase configuration.
* @param addDependencyJars upload HBase jars and jars for any of the configured
* job classes via the distributed cache (tmpjars).
*
* @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should
* have write permissions to this directory, and this should not be a subdirectory of rootdir.
* After the job is finished, restore directory can be deleted.
* @param splitAlgo algorithm to split
* @param numSplitsPerRegion how many input splits to generate per one region
* @throws IOException When setting up the details fails.
* @see TableSnapshotInputFormat
*/
public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,
Class<? extends TableMapper> mapper,
Class<?> outputKeyClass,
Class<?> outputValueClass, Job job,
boolean addDependencyJars, Path tmpRestoreDir,
RegionSplitter.SplitAlgorithm splitAlgo,
int numSplitsPerRegion)
throws IOException {
TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir, splitAlgo,
numSplitsPerRegion);
initTableMapperJob(snapshotName, scan, mapper, outputKeyClass,
outputValueClass, job, addDependencyJars, false, TableSnapshotInputFormat.class);
resetCacheConfig(job.getConfiguration());
}
/**
* Use this before submitting a Multi TableMap job. It will appropriately set
* up the job.

24
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableSnapshotInputFormat.java
View File

@ -33,7 +33,6 @@ import org.apache.hadoop.hbase.client.Result;
import org.apache.hadoop.hbase.client.Scan;
import org.apache.hadoop.hbase.client.metrics.ScanMetrics;
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapreduce.InputFormat;
import org.apache.hadoop.mapreduce.InputSplit;
@ -67,10 +66,8 @@ import org.apache.hadoop.hbase.shaded.com.google.common.annotations.VisibleForTe
* }
* </pre>
* <p>
* Internally, this input format restores the snapshot into the given tmp directory. By default,
* and similar to {@link TableInputFormat} an InputSplit is created per region, but optionally you
* can run N mapper tasks per every region, in which case the region key range will be split to
* N sub-ranges and an InputSplit will be created per sub-range. The region is opened for reading
* Internally, this input format restores the snapshot into the given tmp directory. Similar to
* {@link TableInputFormat} an InputSplit is created per region. The region is opened for reading
* from each RecordReader. An internal RegionScanner is used to execute the
* {@link org.apache.hadoop.hbase.CellScanner} obtained from the user.
* <p>
@ -219,21 +216,4 @@ public class TableSnapshotInputFormat extends InputFormat<ImmutableBytesWritable
throws IOException {
TableSnapshotInputFormatImpl.setInput(job.getConfiguration(), snapshotName, restoreDir);
}
/**
* Configures the job to use TableSnapshotInputFormat to read from a snapshot.
* @param job the job to configure
* @param snapshotName the name of the snapshot to read from
* @param restoreDir a temporary directory to restore the snapshot into. Current user should
* have write permissions to this directory, and this should not be a subdirectory of rootdir.
* After the job is finished, restoreDir can be deleted.
* @param splitAlgo split algorithm to generate splits from region
* @param numSplitsPerRegion how many input splits to generate per one region
* @throws IOException if an error occurs
*/
public static void setInput(Job job, String snapshotName, Path restoreDir,
RegionSplitter.SplitAlgorithm splitAlgo, int numSplitsPerRegion) throws IOException {
TableSnapshotInputFormatImpl.setInput(job.getConfiguration(), snapshotName, restoreDir,
splitAlgo, numSplitsPerRegion);
}
}

93
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/TableSnapshotInputFormatImpl.java
View File

@ -45,7 +45,6 @@ import org.apache.hadoop.hbase.snapshot.SnapshotDescriptionUtils;
import org.apache.hadoop.hbase.snapshot.SnapshotManifest;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.FSUtils;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.io.Writable;
import java.io.ByteArrayOutputStream;
@ -75,17 +74,6 @@ public class TableSnapshotInputFormatImpl {
"hbase.tablesnapshotinputformat.locality.cutoff.multiplier";
private static final float DEFAULT_LOCALITY_CUTOFF_MULTIPLIER = 0.8f;
/**
* For MapReduce jobs running multiple mappers per region, determines
* what split algorithm we should be using to find split points for scanners.
*/
public static final String SPLIT_ALGO = "hbase.mapreduce.split.algorithm";
/**
* For MapReduce jobs running multiple mappers per region, determines
* number of splits to generate per region.
*/
public static final String NUM_SPLITS_PER_REGION = "hbase.mapreduce.splits.per.region";
/**
* Implementation class for InputSplit logic common between mapred and mapreduce.
*/
@ -273,33 +261,9 @@ public class TableSnapshotInputFormatImpl {
// the temp dir where the snapshot is restored
Path restoreDir = new Path(conf.get(RESTORE_DIR_KEY));
RegionSplitter.SplitAlgorithm splitAlgo = getSplitAlgo(conf);
int numSplits = conf.getInt(NUM_SPLITS_PER_REGION, 1);
return getSplits(scan, manifest, regionInfos, restoreDir, conf, splitAlgo, numSplits);
return getSplits(scan, manifest, regionInfos, restoreDir, conf);
}
public static RegionSplitter.SplitAlgorithm getSplitAlgo(Configuration conf) throws IOException{
String splitAlgoClassName = conf.get(SPLIT_ALGO);
if (splitAlgoClassName == null)
return null;
try {
return ((Class<? extends RegionSplitter.SplitAlgorithm>)
Class.forName(splitAlgoClassName)).newInstance();
} catch (ClassNotFoundException e) {
throw new IOException("SplitAlgo class " + splitAlgoClassName +
" is not found", e);
} catch (InstantiationException e) {
throw new IOException("SplitAlgo class " + splitAlgoClassName +
" is not instantiable", e);
} catch (IllegalAccessException e) {
throw new IOException("SplitAlgo class " + splitAlgoClassName +
" is not instantiable", e);
}
}
public static List<HRegionInfo> getRegionInfosFromManifest(SnapshotManifest manifest) {
List<SnapshotRegionManifest> regionManifests = manifest.getRegionManifests();
if (regionManifests == null) {
@ -344,12 +308,6 @@ public class TableSnapshotInputFormatImpl {
public static List<InputSplit> getSplits(Scan scan, SnapshotManifest manifest,
List<HRegionInfo> regionManifests, Path restoreDir, Configuration conf) throws IOException {
return getSplits(scan, manifest, regionManifests, restoreDir, conf, null, 1);
}
public static List<InputSplit> getSplits(Scan scan, SnapshotManifest manifest,
List<HRegionInfo> regionManifests, Path restoreDir,
Configuration conf, RegionSplitter.SplitAlgorithm sa, int numSplits) throws IOException {
// load table descriptor
TableDescriptor htd = manifest.getTableDescriptor();
@ -359,25 +317,6 @@ public class TableSnapshotInputFormatImpl {
for (HRegionInfo hri : regionManifests) {
// load region descriptor
if (numSplits > 1) {
byte[][] sp = sa.split(hri.getStartKey(), hri.getEndKey(), numSplits, true);
for (int i = 0; i < sp.length - 1; i++) {
if (CellUtil.overlappingKeys(scan.getStartRow(), scan.getStopRow(), sp[i],
sp[i + 1])) {
// compute HDFS locations from snapshot files (which will get the locations for
// referred hfiles)
List<String> hosts = getBestLocations(conf,
HRegion.computeHDFSBlocksDistribution(conf, htd, hri, tableDir));
int len = Math.min(3, hosts.size());
hosts = hosts.subList(0, len);
Scan boundedScan = new Scan(scan);
boundedScan.setStartRow(sp[i]);
boundedScan.setStopRow(sp[i + 1]);
splits.add(new InputSplit(htd, hri, hosts, boundedScan, restoreDir));
}
}
} else {
if (CellUtil.overlappingKeys(scan.getStartRow(), scan.getStopRow(), hri.getStartKey(),
hri.getEndKey())) {
// compute HDFS locations from snapshot files (which will get the locations for
@ -390,7 +329,6 @@ public class TableSnapshotInputFormatImpl {
splits.add(new InputSplit(htd, hri, hosts, scan, restoreDir));
}
}
}
return splits;
@ -457,35 +395,8 @@ public class TableSnapshotInputFormatImpl {
*/
public static void setInput(Configuration conf, String snapshotName, Path restoreDir)
throws IOException {
setInput(conf, snapshotName, restoreDir, null, 1);
}
/**
* Configures the job to use TableSnapshotInputFormat to read from a snapshot.
* @param conf the job to configure
* @param snapshotName the name of the snapshot to read from
* @param restoreDir a temporary directory to restore the snapshot into. Current user should
* have write permissions to this directory, and this should not be a subdirectory of rootdir.
* After the job is finished, restoreDir can be deleted.
* @param numSplitsPerRegion how many input splits to generate per one region
* @param splitAlgo SplitAlgorithm to be used when generating InputSplits
* @throws IOException if an error occurs
*/
public static void setInput(Configuration conf, String snapshotName, Path restoreDir,
RegionSplitter.SplitAlgorithm splitAlgo, int numSplitsPerRegion)
throws IOException {
conf.set(SNAPSHOT_NAME_KEY, snapshotName);
if (numSplitsPerRegion < 1) {
throw new IllegalArgumentException("numSplits must be >= 1, " +
"illegal numSplits : " + numSplitsPerRegion);
}
if (splitAlgo == null && numSplitsPerRegion > 1) {
throw new IllegalArgumentException("Split algo can't be null when numSplits > 1");
}
if (splitAlgo != null) {
conf.set(SPLIT_ALGO, splitAlgo.getClass().getName());
}
conf.setInt(NUM_SPLITS_PER_REGION, numSplitsPerRegion);
Path rootDir = FSUtils.getRootDir(conf);
FileSystem fs = rootDir.getFileSystem(conf);

29
hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapred/TestTableSnapshotInputFormat.java
View File

@ -30,7 +30,6 @@ import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
import org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormatTestBase;
import org.apache.hadoop.hbase.testclassification.VerySlowMapReduceTests;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.RegionSplitter;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.mapred.InputSplit;
import org.apache.hadoop.mapred.JobClient;
@ -138,20 +137,20 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
@Test
@Override
public void testWithMockedMapReduceMultiRegion() throws Exception {
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceMultiRegion", 10, 1, 10);
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceMultiRegion", 10, 10);
}
@Test
@Override
public void testWithMapReduceMultiRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceMultiRegion", 10, 1, 10, false);
testWithMapReduce(UTIL, "testWithMapReduceMultiRegion", 10, 10, false);
}
@Test
@Override
// run the MR job while HBase is offline
public void testWithMapReduceAndOfflineHBaseMultiRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceAndOfflineHBaseMultiRegion", 10, 1, 10, true);
testWithMapReduce(UTIL, "testWithMapReduceAndOfflineHBaseMultiRegion", 10, 10, true);
}
@Override
@ -165,7 +164,7 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
@Override
protected void testWithMockedMapReduce(HBaseTestingUtility util, String snapshotName,
int numRegions, int numSplitsPerRegion, int expectedNumSplits) throws Exception {
int numRegions, int expectedNumSplits) throws Exception {
setupCluster();
final TableName tableName = TableName.valueOf(name.getMethodName());
try {
@ -175,16 +174,9 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
JobConf job = new JobConf(util.getConfiguration());
Path tmpTableDir = util.getDataTestDirOnTestFS(snapshotName);
if (numSplitsPerRegion > 1) {
TableMapReduceUtil.initTableSnapshotMapJob(snapshotName,
COLUMNS, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, false, tmpTableDir, new RegionSplitter.UniformSplit(),
numSplitsPerRegion);
} else {
TableMapReduceUtil.initTableSnapshotMapJob(snapshotName,
COLUMNS, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, false, tmpTableDir);
}
// mapred doesn't support start and end keys? o.O
verifyWithMockedMapReduce(job, numRegions, expectedNumSplits, getStartRow(), getEndRow());
@ -233,16 +225,16 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
@Override
protected void testWithMapReduceImpl(HBaseTestingUtility util, TableName tableName,
String snapshotName, Path tableDir, int numRegions, int numSplitsPerRegion, int expectedNumSplits,
String snapshotName, Path tableDir, int numRegions, int expectedNumSplits,
boolean shutdownCluster) throws Exception {
doTestWithMapReduce(util, tableName, snapshotName, getStartRow(), getEndRow(), tableDir,
numRegions, numSplitsPerRegion, expectedNumSplits, shutdownCluster);
numRegions, expectedNumSplits, shutdownCluster);
}
// this is also called by the IntegrationTestTableSnapshotInputFormat
public static void doTestWithMapReduce(HBaseTestingUtility util, TableName tableName,
String snapshotName, byte[] startRow, byte[] endRow, Path tableDir, int numRegions,
int numSplitsPerRegion,int expectedNumSplits, boolean shutdownCluster) throws Exception {
int expectedNumSplits, boolean shutdownCluster) throws Exception {
//create the table and snapshot
createTableAndSnapshot(util, tableName, snapshotName, startRow, endRow, numRegions);
@ -259,16 +251,9 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addDependencyJarsForClasses(jobConf,
TestTableSnapshotInputFormat.class);
if(numSplitsPerRegion > 1) {
TableMapReduceUtil.initTableSnapshotMapJob(snapshotName, COLUMNS,
TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, jobConf, true, tableDir, new RegionSplitter.UniformSplit(),
numSplitsPerRegion);
} else {
TableMapReduceUtil.initTableSnapshotMapJob(snapshotName, COLUMNS,
TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, jobConf, true, tableDir);
}
jobConf.setReducerClass(TestTableSnapshotInputFormat.TestTableSnapshotReducer.class);
jobConf.setNumReduceTasks(1);

23
hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TableSnapshotInputFormatTestBase.java
View File

@ -78,10 +78,10 @@ public abstract class TableSnapshotInputFormatTestBase {
}
protected abstract void testWithMockedMapReduce(HBaseTestingUtility util, String snapshotName,
int numRegions, int numSplitsPerRegion, int expectedNumSplits) throws Exception;
int numRegions, int expectedNumSplits) throws Exception;
protected abstract void testWithMapReduceImpl(HBaseTestingUtility util, TableName tableName,
String snapshotName, Path tableDir, int numRegions, int numSplitsPerRegion, int expectedNumSplits,
String snapshotName, Path tableDir, int numRegions, int expectedNumSplits,
boolean shutdownCluster) throws Exception;
protected abstract byte[] getStartRow();
@ -90,33 +90,28 @@ public abstract class TableSnapshotInputFormatTestBase {
@Test
public void testWithMockedMapReduceSingleRegion() throws Exception {
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceSingleRegion", 1, 1, 1);
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceSingleRegion", 1, 1);
}
@Test
public void testWithMockedMapReduceMultiRegion() throws Exception {
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceMultiRegion", 10, 1, 8);
testWithMockedMapReduce(UTIL, "testWithMockedMapReduceMultiRegion", 10, 8);
}
@Test
public void testWithMapReduceSingleRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceSingleRegion", 1, 1, 1, false);
testWithMapReduce(UTIL, "testWithMapReduceSingleRegion", 1, 1, false);
}
@Test
public void testWithMapReduceMultiRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceMultiRegion", 10, 1, 8, false);
}
@Test
public void testWithMapReduceMultipleMappersPerRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceMultiRegion", 10, 5, 50, false);
testWithMapReduce(UTIL, "testWithMapReduceMultiRegion", 10, 8, false);
}
@Test
// run the MR job while HBase is offline
public void testWithMapReduceAndOfflineHBaseMultiRegion() throws Exception {
testWithMapReduce(UTIL, "testWithMapReduceAndOfflineHBaseMultiRegion", 10, 1, 8, true);
testWithMapReduce(UTIL, "testWithMapReduceAndOfflineHBaseMultiRegion", 10, 8, true);
}
// Test that snapshot restore does not create back references in the HBase root dir.
@ -164,13 +159,13 @@ public abstract class TableSnapshotInputFormatTestBase {
String snapshotName, Path tmpTableDir) throws Exception;
protected void testWithMapReduce(HBaseTestingUtility util, String snapshotName,
int numRegions, int numSplitsPerRegion, int expectedNumSplits, boolean shutdownCluster) throws Exception {
int numRegions, int expectedNumSplits, boolean shutdownCluster) throws Exception {
setupCluster();
try {
Path tableDir = util.getDataTestDirOnTestFS(snapshotName);
TableName tableName = TableName.valueOf("testWithMapReduce");
testWithMapReduceImpl(util, tableName, snapshotName, tableDir, numRegions,
numSplitsPerRegion, expectedNumSplits, shutdownCluster);
expectedNumSplits, shutdownCluster);
} finally {
tearDownCluster();
}

25
hbase-mapreduce/src/test/java/org/apache/hadoop/hbase/mapreduce/TestTableSnapshotInputFormat.java
View File

@ -64,7 +64,6 @@ import org.apache.hadoop.hbase.client.Admin;
import org.apache.hadoop.hbase.client.Table;
import org.apache.hadoop.hbase.snapshot.SnapshotTestingUtils;
import org.apache.hadoop.hbase.util.FSUtils;
import org.apache.hadoop.hbase.util.RegionSplitter;
@Category({VerySlowMapReduceTests.class, LargeTests.class})
public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBase {
@ -210,7 +209,7 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
@Override
public void testWithMockedMapReduce(HBaseTestingUtility util, String snapshotName,
int numRegions, int numSplitsPerRegion, int expectedNumSplits) throws Exception {
int numRegions, int expectedNumSplits) throws Exception {
setupCluster();
final TableName tableName = TableName.valueOf(name.getMethodName());
try {
@ -221,16 +220,9 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
Path tmpTableDir = util.getDataTestDirOnTestFS(snapshotName);
Scan scan = new Scan(getStartRow(), getEndRow()); // limit the scan
if (numSplitsPerRegion > 1) {
TableMapReduceUtil.initTableSnapshotMapperJob(snapshotName,
scan, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, false, tmpTableDir, new RegionSplitter.UniformSplit(),
numSplitsPerRegion);
} else {
TableMapReduceUtil.initTableSnapshotMapperJob(snapshotName,
scan, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, false, tmpTableDir);
}
verifyWithMockedMapReduce(job, numRegions, expectedNumSplits, getStartRow(), getEndRow());
@ -332,16 +324,16 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
@Override
protected void testWithMapReduceImpl(HBaseTestingUtility util, TableName tableName,
String snapshotName, Path tableDir, int numRegions, int numSplitsPerRegion,
int expectedNumSplits, boolean shutdownCluster) throws Exception {
String snapshotName, Path tableDir, int numRegions, int expectedNumSplits,
boolean shutdownCluster) throws Exception {
doTestWithMapReduce(util, tableName, snapshotName, getStartRow(), getEndRow(), tableDir,
numRegions, numSplitsPerRegion, expectedNumSplits, shutdownCluster);
numRegions, expectedNumSplits, shutdownCluster);
}
// this is also called by the IntegrationTestTableSnapshotInputFormat
public static void doTestWithMapReduce(HBaseTestingUtility util, TableName tableName,
String snapshotName, byte[] startRow, byte[] endRow, Path tableDir, int numRegions,
int numSplitsPerRegion, int expectedNumSplits, boolean shutdownCluster) throws Exception {
int expectedNumSplits, boolean shutdownCluster) throws Exception {
LOG.info("testing with MapReduce");
@ -362,16 +354,9 @@ public class TestTableSnapshotInputFormat extends TableSnapshotInputFormatTestBa
TableMapReduceUtil.addDependencyJarsForClasses(job.getConfiguration(),
TestTableSnapshotInputFormat.class);
if (numSplitsPerRegion > 1) {
TableMapReduceUtil.initTableSnapshotMapperJob(snapshotName,
scan, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, true, tableDir, new RegionSplitter.UniformSplit(),
numSplitsPerRegion);
} else {
TableMapReduceUtil.initTableSnapshotMapperJob(snapshotName,
scan, TestTableSnapshotMapper.class, ImmutableBytesWritable.class,
NullWritable.class, job, true, tableDir);
}
job.setReducerClass(TestTableSnapshotInputFormat.TestTableSnapshotReducer.class);
job.setNumReduceTasks(1);

2
hbase-server/src/main/java/org/apache/hadoop/hbase/client/ClientSideRegionScanner.java
View File

@ -53,8 +53,6 @@ public class ClientSideRegionScanner extends AbstractClientScanner {
// region is immutable, set isolation level
scan.setIsolationLevel(IsolationLevel.READ_UNCOMMITTED);
htd = TableDescriptorBuilder.newBuilder(htd).setReadOnly(true).build();
// open region from the snapshot directory
this.region = HRegion.openHRegion(conf, fs, rootDir, hri, htd, null, null, null);

71
hbase-server/src/main/java/org/apache/hadoop/hbase/util/RegionSplitter.java
View File

@ -49,7 +49,6 @@ import org.apache.hadoop.hbase.client.TableDescriptor;
import org.apache.hadoop.hbase.ClusterStatus;
import org.apache.hadoop.hbase.ClusterStatus.Option;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.HRegionInfo;
import org.apache.hadoop.hbase.HRegionLocation;
import org.apache.hadoop.hbase.MetaTableAccessor;
@ -182,17 +181,6 @@ public class RegionSplitter {
*/
byte[][] split(int numRegions);
/**
* Some MapReduce jobs may want to run multiple mappers per region,
* this is intended for such usecase.
*
* @param start first row (inclusive)
* @param end last row (exclusive)
* @param numSplits number of splits to generate
* @param inclusive whether start and end are returned as split points
*/
byte[][] split(byte[] start, byte[] end, int numSplits, boolean inclusive);
/**
* In HBase, the first row is represented by an empty byte array. This might
* cause problems with your split algorithm or row printing. All your APIs
@ -930,39 +918,6 @@ public class RegionSplitter {
return convertToBytes(splits);
}
@Override
public byte[][] split(byte[] start, byte[] end, int numSplits, boolean inclusive) {
BigInteger s = convertToBigInteger(start);
BigInteger e = convertToBigInteger(end);
Preconditions.checkArgument(e.compareTo(s) > 0,
"last row (%s) is configured less than first row (%s)", rowToStr(end),
end);
// +1 to range because the last row is inclusive
BigInteger range = e.subtract(s).add(BigInteger.ONE);
Preconditions.checkState(range.compareTo(BigInteger.valueOf(numSplits)) >= 0,
"split granularity (%s) is greater than the range (%s)", numSplits, range);
BigInteger[] splits = new BigInteger[numSplits - 1];
BigInteger sizeOfEachSplit = range.divide(BigInteger.valueOf(numSplits));
for (int i = 1; i < numSplits; i++) {
// NOTE: this means the last region gets all the slop.
// This is not a big deal if we're assuming n << MAXHEX
splits[i - 1] = s.add(sizeOfEachSplit.multiply(BigInteger
.valueOf(i)));
}
if (inclusive) {
BigInteger[] inclusiveSplitPoints = new BigInteger[numSplits + 1];
inclusiveSplitPoints[0] = convertToBigInteger(start);
inclusiveSplitPoints[numSplits] = convertToBigInteger(end);
System.arraycopy(splits, 0, inclusiveSplitPoints, 1, splits.length);
return convertToBytes(inclusiveSplitPoints);
} else {
return convertToBytes(splits);
}
}
public byte[] firstRow() {
return convertToByte(firstRowInt);
}
@ -1106,32 +1061,6 @@ public class RegionSplitter {
return splits == null? null: Arrays.copyOfRange(splits, 1, splits.length - 1);
}
@Override
public byte[][] split(byte[] start, byte[] end, int numSplits, boolean inclusive) {
if (Arrays.equals(start, HConstants.EMPTY_BYTE_ARRAY)) {
start = firstRowBytes;
}
if (Arrays.equals(end, HConstants.EMPTY_BYTE_ARRAY)) {
end = lastRowBytes;
}
Preconditions.checkArgument(
Bytes.compareTo(end, start) > 0,
"last row (%s) is configured less than first row (%s)",
Bytes.toStringBinary(end),
Bytes.toStringBinary(start));
byte[][] splits = Bytes.split(start, end, true,
numSplits - 1);
Preconditions.checkState(splits != null,
"Could not calculate input splits with given user input: " + this);
if (inclusive) {
return splits;
} else {
// remove endpoints, which are included in the splits list
return Arrays.copyOfRange(splits, 1, splits.length - 1);
}
}
@Override
public byte[] firstRow() {
return firstRowBytes;

20
hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestRegionSplitter.java
View File

@ -168,16 +168,6 @@ public class TestRegionSplitter {
// Halfway between df... and ff... should be ef....
splitPoint = splitter.split("dfffffff".getBytes(), lastRow);
assertArrayEquals(splitPoint,"efffffff".getBytes());
// Check splitting region with multiple mappers per region
byte[][] splits = splitter.split("00000000".getBytes(), "30000000".getBytes(), 3, false);
assertEquals(2, splits.length);
assertArrayEquals(splits[0], "10000000".getBytes());
assertArrayEquals(splits[1], "20000000".getBytes());
splits = splitter.split("00000000".getBytes(), "20000000".getBytes(), 2, true);
assertEquals(3, splits.length);
assertArrayEquals(splits[1], "10000000".getBytes());
}
/**
@ -225,16 +215,6 @@ public class TestRegionSplitter {
splitPoint = splitter.split(new byte[] {'a', 'a', 'a'}, new byte[] {'a', 'a', 'b'});
assertArrayEquals(splitPoint, new byte[] { 'a', 'a', 'a', (byte) 0x80 });
// Check splitting region with multiple mappers per region
byte[][] splits = splitter.split(new byte[] {'a', 'a', 'a'}, new byte[] {'a', 'a', 'd'}, 3, false);
assertEquals(2, splits.length);
assertArrayEquals(splits[0], new byte[]{'a', 'a', 'b'});
assertArrayEquals(splits[1], new byte[]{'a', 'a', 'c'});
splits = splitter.split(new byte[] {'a', 'a', 'a'}, new byte[] {'a', 'a', 'e'}, 2, true);
assertEquals(3, splits.length);
assertArrayEquals(splits[1], new byte[] { 'a', 'a', 'c'});
}
@Test