HBASE-20940 HStore.cansplit should not allow split to happen if it has references (Vishal Khandelwal)

This commit is contained in:
Andrew Purtell 2018-08-17 15:01:44 -07:00
parent 9da5d3a481
commit 52ba33aa73
6 changed files with 210 additions and 27 deletions

View File

@ -1659,7 +1659,17 @@ public class HStore implements Store, HeapSize, StoreConfigInformation, Propagat
@Override
public boolean hasReferences() {
return StoreUtils.hasReferences(this.storeEngine.getStoreFileManager().getStorefiles());
List<HStoreFile> reloadedStoreFiles = null;
try {
// Reloading the store files from file system due to HBASE-20940. As split can happen with an
// region which has references
reloadedStoreFiles = loadStoreFiles();
return StoreUtils.hasReferences(reloadedStoreFiles);
} catch (IOException ioe) {
LOG.error("Error trying to determine if store has references, assuming references exists",
ioe);
return true;
}
}
/**

View File

@ -21,7 +21,6 @@ import static org.apache.hadoop.hbase.HConstants.HBASE_CLIENT_META_OPERATION_TIM
import static org.apache.hadoop.hbase.master.LoadBalancer.TABLES_ON_MASTER;
import static org.junit.Assert.assertEquals;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
@ -35,18 +34,21 @@ import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicBoolean;
import java.util.stream.Collectors;
import java.util.stream.IntStream;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.hbase.HBaseClassTestRule;
import org.apache.hadoop.hbase.HBaseTestingUtility;
import org.apache.hadoop.hbase.MemoryCompactionPolicy;
import org.apache.hadoop.hbase.ServerName;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.Waiter.ExplainingPredicate;
import org.apache.hadoop.hbase.io.ByteBufferPool;
import org.apache.hadoop.hbase.regionserver.CompactingMemStore;
import org.apache.hadoop.hbase.regionserver.HRegion;
import org.apache.hadoop.hbase.testclassification.ClientTests;
import org.apache.hadoop.hbase.testclassification.LargeTests;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.RetryCounter;
import org.apache.hadoop.hbase.util.Threads;
import org.junit.AfterClass;
import org.junit.BeforeClass;
@ -124,7 +126,7 @@ public class TestAsyncTableGetMultiThreaded {
}
@Test
public void test() throws IOException, InterruptedException, ExecutionException {
public void test() throws Exception {
int numThreads = 20;
AtomicBoolean stop = new AtomicBoolean(false);
ExecutorService executor =
@ -137,9 +139,31 @@ public class TestAsyncTableGetMultiThreaded {
Collections.shuffle(Arrays.asList(SPLIT_KEYS), new Random(123));
Admin admin = TEST_UTIL.getAdmin();
for (byte[] splitPoint : SPLIT_KEYS) {
int oldRegionCount = admin.getRegions(TABLE_NAME).size();
admin.split(TABLE_NAME, splitPoint);
TEST_UTIL.waitFor(30000, new ExplainingPredicate<Exception>() {
@Override
public boolean evaluate() throws Exception {
return TEST_UTIL.getMiniHBaseCluster().getRegions(TABLE_NAME).size() > oldRegionCount;
}
@Override
public String explainFailure() throws Exception {
return "Split has not finished yet";
}
});
for (HRegion region : TEST_UTIL.getHBaseCluster().getRegions(TABLE_NAME)) {
region.compact(true);
//Waiting for compaction to complete and references are cleaned up
RetryCounter retrier = new RetryCounter(30, 1, TimeUnit.SECONDS);
while (CompactionState.NONE != admin
.getCompactionStateForRegion(region.getRegionInfo().getRegionName())
&& retrier.shouldRetry()) {
retrier.sleepUntilNextRetry();
}
region.getStores().get(0).closeAndArchiveCompactedFiles();
}
Thread.sleep(5000);
admin.balance(true);

View File

@ -107,6 +107,10 @@ public class TestChangingEncoding {
public static void setUpBeforeClass() throws Exception {
// Use a small flush size to create more HFiles.
conf.setInt(HConstants.HREGION_MEMSTORE_FLUSH_SIZE, 1024 * 1024);
// Disabling split to make sure split does not cause modify column to wait which timesout test
// sometime
conf.set(HConstants.HBASE_REGION_SPLIT_POLICY_KEY,
"org.apache.hadoop.hbase.regionserver.DisabledRegionSplitPolicy");
// ((Log4JLogger)RpcServerImplementation.LOG).getLogger().setLevel(Level.TRACE);
// ((Log4JLogger)RpcClient.LOG).getLogger().setLevel(Level.TRACE);
TEST_UTIL.startMiniCluster();

View File

@ -48,7 +48,9 @@ import org.apache.hadoop.hbase.MiniHBaseCluster;
import org.apache.hadoop.hbase.NamespaceDescriptor;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.Waiter;
import org.apache.hadoop.hbase.Waiter.ExplainingPredicate;
import org.apache.hadoop.hbase.client.Admin;
import org.apache.hadoop.hbase.client.CompactionState;
import org.apache.hadoop.hbase.client.Connection;
import org.apache.hadoop.hbase.client.ConnectionFactory;
import org.apache.hadoop.hbase.client.DoNotRetryRegionException;
@ -81,6 +83,7 @@ import org.apache.hadoop.hbase.snapshot.RestoreSnapshotException;
import org.apache.hadoop.hbase.testclassification.MediumTests;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.FSUtils;
import org.apache.hadoop.hbase.util.RetryCounter;
import org.apache.hadoop.hbase.util.Threads;
import org.apache.zookeeper.KeeperException;
import org.junit.After;
@ -365,6 +368,23 @@ public class TestNamespaceAuditor {
HRegion regionToSplit = UTIL.getMiniHBaseCluster().getRegions(tableTwo).stream()
.filter(r -> r.getRegionInfo().containsRow(splitKey)).findFirst().get();
regionToSplit.compact(true);
// Waiting for compaction to finish
UTIL.waitFor(30000, new Waiter.Predicate<Exception>() {
@Override
public boolean evaluate() throws Exception {
return (CompactionState.NONE == ADMIN
.getCompactionStateForRegion(regionToSplit.getRegionInfo().getRegionName()));
}
});
// Cleaning compacted references for split to proceed
regionToSplit.getStores().stream().forEach(s -> {
try {
s.closeAndArchiveCompactedFiles();
} catch (IOException e1) {
LOG.error("Error whiling cleaning compacted file");
}
});
// the above compact may quit immediately if there is a compaction ongoing, so here we need to
// wait a while to let the ongoing compaction finish.
UTIL.waitFor(10000, regionToSplit::isSplittable);

View File

@ -24,10 +24,13 @@ import static org.junit.Assert.assertTrue;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.Random;
import java.util.Set;
import java.util.TreeSet;
import java.util.concurrent.TimeUnit;
import java.util.stream.Collectors;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.ChoreService;
@ -41,6 +44,8 @@ import org.apache.hadoop.hbase.ScheduledChore;
import org.apache.hadoop.hbase.Stoppable;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.Admin;
import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;
import org.apache.hadoop.hbase.client.CompactionState;
import org.apache.hadoop.hbase.client.Connection;
import org.apache.hadoop.hbase.client.ConnectionFactory;
import org.apache.hadoop.hbase.client.Get;
@ -49,13 +54,19 @@ import org.apache.hadoop.hbase.client.RegionInfo;
import org.apache.hadoop.hbase.client.RegionLocator;
import org.apache.hadoop.hbase.client.Result;
import org.apache.hadoop.hbase.client.Table;
import org.apache.hadoop.hbase.client.TableDescriptor;
import org.apache.hadoop.hbase.client.TableDescriptorBuilder;
import org.apache.hadoop.hbase.testclassification.LargeTests;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.hadoop.hbase.util.Pair;
import org.apache.hadoop.hbase.util.PairOfSameType;
import org.apache.hadoop.hbase.util.RetryCounter;
import org.apache.hadoop.hbase.util.StoppableImplementation;
import org.apache.hadoop.hbase.util.Threads;
import org.apache.hbase.thirdparty.com.google.common.collect.Iterators;
import org.apache.hbase.thirdparty.com.google.common.collect.Maps;
import org.junit.AfterClass;
import org.junit.Assert;
import org.junit.BeforeClass;
import org.junit.ClassRule;
import org.junit.Rule;
@ -65,8 +76,6 @@ import org.junit.rules.TestName;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.apache.hbase.thirdparty.com.google.common.collect.Iterators;
@Category(LargeTests.class)
public class TestEndToEndSplitTransaction {
@ -92,6 +101,78 @@ public class TestEndToEndSplitTransaction {
TEST_UTIL.shutdownMiniCluster();
}
/*
* This is the test for : HBASE-20940 This test will split the region and try to open an reference
* over store file. Once store file has any reference, it makes sure that region can't be split
* @throws Exception
*/
@Test
public void testCanSplitJustAfterASplit() throws Exception {
LOG.info("Starting testCanSplitJustAfterASplit");
byte[] fam = Bytes.toBytes("cf_split");
TableName tableName = TableName.valueOf("CanSplitTable");
Table source = TEST_UTIL.getConnection().getTable(tableName);
Admin admin = TEST_UTIL.getAdmin();
Map<String, StoreFileReader> scanner = Maps.newHashMap();
try {
TableDescriptor htd = TableDescriptorBuilder.newBuilder(tableName)
.setColumnFamily(ColumnFamilyDescriptorBuilder.of(fam)).build();
admin.createTable(htd);
TEST_UTIL.loadTable(source, fam);
List<HRegion> regions = TEST_UTIL.getHBaseCluster().getRegions(tableName);
regions.get(0).forceSplit(null);
admin.split(tableName);
while (regions.size() <= 1) {
regions = TEST_UTIL.getHBaseCluster().getRegions(tableName);
regions.stream()
.forEach(r -> r.getStores().get(0).getStorefiles().stream()
.filter(
s -> s.isReference() && !scanner.containsKey(r.getRegionInfo().getEncodedName()))
.forEach(sf -> {
StoreFileReader reader = ((HStoreFile) sf).getReader();
reader.getStoreFileScanner(true, false, false, 0, 0, false);
scanner.put(r.getRegionInfo().getEncodedName(), reader);
LOG.info("Got reference to file = " + sf.getPath() + ",for region = "
+ r.getRegionInfo().getEncodedName());
}));
}
Assert.assertTrue("Regions did not split properly", regions.size() > 1);
Assert.assertTrue("Could not get reference any of the store file", scanner.size() > 1);
RetryCounter retrier = new RetryCounter(30, 1, TimeUnit.SECONDS);
while (CompactionState.NONE != admin.getCompactionState(tableName) && retrier.shouldRetry()) {
retrier.sleepUntilNextRetry();
}
Assert.assertEquals("Compaction did not complete in 30 secs", CompactionState.NONE,
admin.getCompactionState(tableName));
regions.stream()
.filter(region -> scanner.containsKey(region.getRegionInfo().getEncodedName()))
.forEach(r -> Assert.assertTrue("Contains an open file reference which can be split",
!r.getStores().get(0).canSplit()));
} finally {
scanner.values().stream().forEach(s -> {
try {
s.close(true);
} catch (IOException ioe) {
LOG.error("Failed while closing store file", ioe);
}
});
scanner.clear();
if (source != null) {
source.close();
}
TEST_UTIL.deleteTableIfAny(tableName);
}
}
/**
* Tests that the client sees meta table changes as atomic during splits
*/
@ -151,18 +232,17 @@ public class TestEndToEndSplitTransaction {
public void run() {
try {
Random random = new Random();
for (int i= 0; i< 5; i++) {
List<RegionInfo> regions =
MetaTableAccessor.getTableRegions(connection, tableName, true);
for (int i = 0; i < 5; i++) {
List<RegionInfo> regions = MetaTableAccessor.getTableRegions(connection, tableName, true);
if (regions.isEmpty()) {
continue;
}
int regionIndex = random.nextInt(regions.size());
//pick a random region and split it into two
// pick a random region and split it into two
RegionInfo region = Iterators.get(regions.iterator(), regionIndex);
//pick the mid split point
// pick the mid split point
int start = 0, end = Integer.MAX_VALUE;
if (region.getStartKey().length > 0) {
start = Bytes.toInt(region.getStartKey());
@ -173,7 +253,7 @@ public class TestEndToEndSplitTransaction {
int mid = start + ((end - start) / 2);
byte[] splitPoint = Bytes.toBytes(mid);
//put some rows to the regions
// put some rows to the regions
addData(start);
addData(mid);
@ -183,11 +263,11 @@ public class TestEndToEndSplitTransaction {
log("Initiating region split for:" + region.getRegionNameAsString());
try {
admin.splitRegion(region.getRegionName(), splitPoint);
//wait until the split is complete
// wait until the split is complete
blockUntilRegionSplit(CONF, 50000, region.getRegionName(), true);
} catch (NotServingRegionException ex) {
//ignore
// ignore
}
}
} catch (Throwable ex) {
@ -226,9 +306,11 @@ public class TestEndToEndSplitTransaction {
/** verify region boundaries obtained from MetaScanner */
void verifyRegionsUsingMetaTableAccessor() throws Exception {
List<RegionInfo> regionList = MetaTableAccessor.getTableRegions(connection, tableName, true);
verifyTableRegions(regionList.stream().collect(Collectors.toCollection(() -> new TreeSet<>(RegionInfo.COMPARATOR))));
verifyTableRegions(regionList.stream()
.collect(Collectors.toCollection(() -> new TreeSet<>(RegionInfo.COMPARATOR))));
regionList = MetaTableAccessor.getAllRegions(connection, true);
verifyTableRegions(regionList.stream().collect(Collectors.toCollection(() -> new TreeSet<>(RegionInfo.COMPARATOR))));
verifyTableRegions(regionList.stream()
.collect(Collectors.toCollection(() -> new TreeSet<>(RegionInfo.COMPARATOR))));
}
/** verify region boundaries obtained from HTable.getStartEndKeys() */
@ -343,7 +425,9 @@ public class TestEndToEndSplitTransaction {
}
}
/** Blocks until the region split is complete in hbase:meta and region server opens the daughters */
/**
* Blocks until the region split is complete in hbase:meta and region server opens the daughters
*/
public static void blockUntilRegionSplit(Configuration conf, long timeout,
final byte[] regionName, boolean waitForDaughters)
throws IOException, InterruptedException {
@ -389,10 +473,32 @@ public class TestEndToEndSplitTransaction {
rem = timeout - (System.currentTimeMillis() - start);
blockUntilRegionIsOpened(conf, rem, daughterB);
// Compacting the new region to make sure references can be cleaned up
compactAndBlockUntilDone(TEST_UTIL.getAdmin(),
TEST_UTIL.getMiniHBaseCluster().getRegionServer(0), daughterA.getRegionName());
compactAndBlockUntilDone(TEST_UTIL.getAdmin(),
TEST_UTIL.getMiniHBaseCluster().getRegionServer(0), daughterB.getRegionName());
removeCompactedFiles(conn, timeout, daughterA);
removeCompactedFiles(conn, timeout, daughterB);
}
}
}
public static void removeCompactedFiles(Connection conn, long timeout, RegionInfo hri)
throws IOException, InterruptedException {
log("remove compacted files for : " + hri.getRegionNameAsString());
List<HRegion> regions = TEST_UTIL.getHBaseCluster().getRegions(hri.getTable());
regions.stream().forEach(r -> {
try {
r.getStores().get(0).closeAndArchiveCompactedFiles();
} catch (IOException ioe) {
LOG.error("failed in removing compacted file", ioe);
}
});
}
public static void blockUntilRegionIsInMeta(Connection conn, long timeout, RegionInfo hri)
throws IOException, InterruptedException {
log("blocking until region is in META: " + hri.getRegionNameAsString());
@ -415,7 +521,9 @@ public class TestEndToEndSplitTransaction {
Table table = conn.getTable(hri.getTable())) {
byte[] row = hri.getStartKey();
// Check for null/empty row. If we find one, use a key that is likely to be in first region.
if (row == null || row.length <= 0) row = new byte[] { '0' };
if (row == null || row.length <= 0) {
row = new byte[] { '0' };
}
Get get = new Get(row);
while (System.currentTimeMillis() - start < timeout) {
try {

View File

@ -31,7 +31,9 @@ import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicBoolean;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
@ -50,6 +52,7 @@ import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.UnknownRegionException;
import org.apache.hadoop.hbase.ZooKeeperConnectionException;
import org.apache.hadoop.hbase.client.Admin;
import org.apache.hadoop.hbase.client.CompactionState;
import org.apache.hadoop.hbase.client.Consistency;
import org.apache.hadoop.hbase.client.Delete;
import org.apache.hadoop.hbase.client.DoNotRetryRegionException;
@ -78,6 +81,10 @@ import org.apache.hadoop.hbase.master.assignment.RegionStates;
import org.apache.hadoop.hbase.procedure2.ProcedureTestingUtility;
import org.apache.hadoop.hbase.regionserver.compactions.CompactionContext;
import org.apache.hadoop.hbase.regionserver.throttle.NoLimitThroughputController;
import org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.RegionStateTransition.TransitionCode;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.ReportRegionStateTransitionRequest;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.ReportRegionStateTransitionResponse;
import org.apache.hadoop.hbase.testclassification.LargeTests;
import org.apache.hadoop.hbase.testclassification.RegionServerTests;
import org.apache.hadoop.hbase.util.Bytes;
@ -85,7 +92,10 @@ import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
import org.apache.hadoop.hbase.util.FSUtils;
import org.apache.hadoop.hbase.util.HBaseFsck;
import org.apache.hadoop.hbase.util.JVMClusterUtil.RegionServerThread;
import org.apache.hadoop.hbase.util.RetryCounter;
import org.apache.hadoop.hbase.util.Threads;
import org.apache.hbase.thirdparty.com.google.protobuf.RpcController;
import org.apache.hbase.thirdparty.com.google.protobuf.ServiceException;
import org.apache.zookeeper.KeeperException;
import org.apache.zookeeper.KeeperException.NodeExistsException;
import org.junit.After;
@ -101,14 +111,6 @@ import org.junit.rules.TestName;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.apache.hbase.thirdparty.com.google.protobuf.RpcController;
import org.apache.hbase.thirdparty.com.google.protobuf.ServiceException;
import org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.RegionStateTransition.TransitionCode;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.ReportRegionStateTransitionRequest;
import org.apache.hadoop.hbase.shaded.protobuf.generated.RegionServerStatusProtos.ReportRegionStateTransitionResponse;
/**
* The below tests are testing split region against a running cluster
*/
@ -386,11 +388,18 @@ public class TestSplitTransactionOnCluster {
// Compact first to ensure we have cleaned up references -- else the split
// will fail.
this.admin.compactRegion(daughter.getRegionName());
RetryCounter retrier = new RetryCounter(30, 1, TimeUnit.SECONDS);
while (CompactionState.NONE != admin.getCompactionStateForRegion(daughter.getRegionName())
&& retrier.shouldRetry()) {
retrier.sleepUntilNextRetry();
}
daughters = cluster.getRegions(tableName);
HRegion daughterRegion = null;
for (HRegion r: daughters) {
for (HRegion r : daughters) {
if (RegionInfo.COMPARATOR.compare(r.getRegionInfo(), daughter) == 0) {
daughterRegion = r;
// Archiving the compacted references file
r.getStores().get(0).closeAndArchiveCompactedFiles();
LOG.info("Found matching HRI: " + daughterRegion);
break;
}
@ -533,11 +542,19 @@ public class TestSplitTransactionOnCluster {
// Call split.
this.admin.splitRegion(hri.getRegionName());
List<HRegion> daughters = checkAndGetDaughters(tableName);
// Before cleanup, get a new master.
HMaster master = abortAndWaitForMaster();
// Now call compact on the daughters and clean up any references.
for (HRegion daughter: daughters) {
for (HRegion daughter : daughters) {
daughter.compact(true);
RetryCounter retrier = new RetryCounter(30, 1, TimeUnit.SECONDS);
while (CompactionState.NONE != admin
.getCompactionStateForRegion(daughter.getRegionInfo().getRegionName())
&& retrier.shouldRetry()) {
retrier.sleepUntilNextRetry();
}
daughter.getStores().get(0).closeAndArchiveCompactedFiles();
assertFalse(daughter.hasReferences());
}
// BUT calling compact on the daughters is not enough. The CatalogJanitor looks