org.apache.hadoop.hbase.regionserver.CompactSplit Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of hbase-server Show documentation
Show all versions of hbase-server Show documentation
Server functionality for HBase
/**
*
* 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.hbase.regionserver;
import java.io.IOException;
import java.io.PrintWriter;
import java.io.StringWriter;
import java.util.Comparator;
import java.util.Iterator;
import java.util.Optional;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.RejectedExecutionException;
import java.util.concurrent.RejectedExecutionHandler;
import java.util.concurrent.ThreadFactory;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.conf.Configuration;
import org.apache.yetus.audience.InterfaceAudience;
import org.apache.hadoop.hbase.conf.ConfigurationManager;
import org.apache.hadoop.hbase.conf.PropagatingConfigurationObserver;
import org.apache.hadoop.hbase.quotas.RegionServerSpaceQuotaManager;
import org.apache.hadoop.hbase.regionserver.compactions.CompactionContext;
import org.apache.hadoop.hbase.regionserver.compactions.CompactionLifeCycleTracker;
import org.apache.hadoop.hbase.regionserver.compactions.CompactionRequest;
import org.apache.hadoop.hbase.regionserver.throttle.CompactionThroughputControllerFactory;
import org.apache.hadoop.hbase.regionserver.throttle.ThroughputController;
import org.apache.hadoop.hbase.security.User;
import org.apache.hadoop.hbase.shaded.com.google.common.annotations.VisibleForTesting;
import org.apache.hadoop.hbase.shaded.com.google.common.base.Preconditions;
import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
import org.apache.hadoop.hbase.util.StealJobQueue;
import org.apache.hadoop.ipc.RemoteException;
import org.apache.hadoop.util.StringUtils;
/**
* Compact region on request and then run split if appropriate
*/
@InterfaceAudience.Private
public class CompactSplit implements PropagatingConfigurationObserver {
private static final Log LOG = LogFactory.getLog(CompactSplit.class);
// Configuration key for the large compaction threads.
public final static String LARGE_COMPACTION_THREADS =
"hbase.regionserver.thread.compaction.large";
public final static int LARGE_COMPACTION_THREADS_DEFAULT = 1;
// Configuration key for the small compaction threads.
public final static String SMALL_COMPACTION_THREADS =
"hbase.regionserver.thread.compaction.small";
public final static int SMALL_COMPACTION_THREADS_DEFAULT = 1;
// Configuration key for split threads
public final static String SPLIT_THREADS = "hbase.regionserver.thread.split";
public final static int SPLIT_THREADS_DEFAULT = 1;
public static final String REGION_SERVER_REGION_SPLIT_LIMIT =
"hbase.regionserver.regionSplitLimit";
public static final int DEFAULT_REGION_SERVER_REGION_SPLIT_LIMIT= 1000;
private final HRegionServer server;
private final Configuration conf;
private final ThreadPoolExecutor longCompactions;
private final ThreadPoolExecutor shortCompactions;
private final ThreadPoolExecutor splits;
private volatile ThroughputController compactionThroughputController;
/**
* Splitting should not take place if the total number of regions exceed this.
* This is not a hard limit to the number of regions but it is a guideline to
* stop splitting after number of online regions is greater than this.
*/
private int regionSplitLimit;
/** @param server */
CompactSplit(HRegionServer server) {
super();
this.server = server;
this.conf = server.getConfiguration();
this.regionSplitLimit = conf.getInt(REGION_SERVER_REGION_SPLIT_LIMIT,
DEFAULT_REGION_SERVER_REGION_SPLIT_LIMIT);
int largeThreads = Math.max(1, conf.getInt(
LARGE_COMPACTION_THREADS, LARGE_COMPACTION_THREADS_DEFAULT));
int smallThreads = conf.getInt(
SMALL_COMPACTION_THREADS, SMALL_COMPACTION_THREADS_DEFAULT);
int splitThreads = conf.getInt(SPLIT_THREADS, SPLIT_THREADS_DEFAULT);
// if we have throttle threads, make sure the user also specified size
Preconditions.checkArgument(largeThreads > 0 && smallThreads > 0);
final String n = Thread.currentThread().getName();
StealJobQueue stealJobQueue = new StealJobQueue(COMPARATOR);
this.longCompactions = new ThreadPoolExecutor(largeThreads, largeThreads,
60, TimeUnit.SECONDS, stealJobQueue,
new ThreadFactory() {
@Override
public Thread newThread(Runnable r) {
String name = n + "-longCompactions-" + System.currentTimeMillis();
return new Thread(r, name);
}
});
this.longCompactions.setRejectedExecutionHandler(new Rejection());
this.longCompactions.prestartAllCoreThreads();
this.shortCompactions = new ThreadPoolExecutor(smallThreads, smallThreads,
60, TimeUnit.SECONDS, stealJobQueue.getStealFromQueue(),
new ThreadFactory() {
@Override
public Thread newThread(Runnable r) {
String name = n + "-shortCompactions-" + System.currentTimeMillis();
return new Thread(r, name);
}
});
this.shortCompactions
.setRejectedExecutionHandler(new Rejection());
this.splits = (ThreadPoolExecutor)
Executors.newFixedThreadPool(splitThreads,
new ThreadFactory() {
@Override
public Thread newThread(Runnable r) {
String name = n + "-splits-" + System.currentTimeMillis();
return new Thread(r, name);
}
});
// compaction throughput controller
this.compactionThroughputController =
CompactionThroughputControllerFactory.create(server, conf);
}
@Override
public String toString() {
return "compaction_queue=("
+ longCompactions.getQueue().size() + ":"
+ shortCompactions.getQueue().size() + ")"
+ ", split_queue=" + splits.getQueue().size();
}
public String dumpQueue() {
StringBuffer queueLists = new StringBuffer();
queueLists.append("Compaction/Split Queue dump:\n");
queueLists.append(" LargeCompation Queue:\n");
BlockingQueue lq = longCompactions.getQueue();
Iterator it = lq.iterator();
while (it.hasNext()) {
queueLists.append(" " + it.next().toString());
queueLists.append("\n");
}
if (shortCompactions != null) {
queueLists.append("\n");
queueLists.append(" SmallCompation Queue:\n");
lq = shortCompactions.getQueue();
it = lq.iterator();
while (it.hasNext()) {
queueLists.append(" " + it.next().toString());
queueLists.append("\n");
}
}
queueLists.append("\n");
queueLists.append(" Split Queue:\n");
lq = splits.getQueue();
it = lq.iterator();
while (it.hasNext()) {
queueLists.append(" " + it.next().toString());
queueLists.append("\n");
}
return queueLists.toString();
}
public synchronized boolean requestSplit(final Region r) {
// don't split regions that are blocking
if (shouldSplitRegion() && ((HRegion)r).getCompactPriority() >= Store.PRIORITY_USER) {
byte[] midKey = ((HRegion)r).checkSplit();
if (midKey != null) {
requestSplit(r, midKey);
return true;
}
}
return false;
}
public synchronized void requestSplit(final Region r, byte[] midKey) {
requestSplit(r, midKey, null);
}
/*
* The User parameter allows the split thread to assume the correct user identity
*/
public synchronized void requestSplit(final Region r, byte[] midKey, User user) {
if (midKey == null) {
LOG.debug("Region " + r.getRegionInfo().getRegionNameAsString() +
" not splittable because midkey=null");
if (((HRegion)r).shouldForceSplit()) {
((HRegion)r).clearSplit();
}
return;
}
try {
this.splits.execute(new SplitRequest(r, midKey, this.server, user));
if (LOG.isDebugEnabled()) {
LOG.debug("Splitting " + r + ", " + this);
}
} catch (RejectedExecutionException ree) {
LOG.info("Could not execute split for " + r, ree);
}
}
public synchronized void requestCompaction(HRegion region, String why, int priority,
CompactionLifeCycleTracker tracker, User user) throws IOException {
requestCompactionInternal(region, why, priority, true, tracker, user);
}
public synchronized void requestCompaction(HRegion region, HStore store, String why, int priority,
CompactionLifeCycleTracker tracker, User user) throws IOException {
requestCompactionInternal(region, store, why, priority, true, tracker, user);
}
private void requestCompactionInternal(HRegion region, String why, int priority,
boolean selectNow, CompactionLifeCycleTracker tracker, User user) throws IOException {
// request compaction on all stores
for (HStore store : region.stores.values()) {
requestCompactionInternal(region, store, why, priority, selectNow, tracker, user);
}
}
private void requestCompactionInternal(HRegion region, HStore store, String why, int priority,
boolean selectNow, CompactionLifeCycleTracker tracker, User user) throws IOException {
if (this.server.isStopped() || (region.getTableDescriptor() != null &&
!region.getTableDescriptor().isCompactionEnabled())) {
return;
}
Optional compaction;
if (selectNow) {
compaction = selectCompaction(region, store, priority, tracker, user);
if (!compaction.isPresent()) {
// message logged inside
return;
}
} else {
compaction = Optional.empty();
}
RegionServerSpaceQuotaManager spaceQuotaManager =
this.server.getRegionServerSpaceQuotaManager();
if (spaceQuotaManager != null &&
spaceQuotaManager.areCompactionsDisabled(region.getTableDescriptor().getTableName())) {
if (LOG.isDebugEnabled()) {
LOG.debug("Ignoring compaction request for " + region +
" as an active space quota violation " + " policy disallows compactions.");
}
return;
}
ThreadPoolExecutor pool;
if (selectNow) {
// compaction.get is safe as we will just return if selectNow is true but no compaction is
// selected
pool = store.throttleCompaction(compaction.get().getRequest().getSize()) ? longCompactions
: shortCompactions;
} else {
// We assume that most compactions are small. So, put system compactions into small
// pool; we will do selection there, and move to large pool if necessary.
pool = shortCompactions;
}
pool.execute(new CompactionRunner(store, region, compaction, pool, user));
region.incrementCompactionsQueuedCount();
if (LOG.isDebugEnabled()) {
String type = (pool == shortCompactions) ? "Small " : "Large ";
LOG.debug(type + "Compaction requested: " + (selectNow ? compaction.toString() : "system")
+ (why != null && !why.isEmpty() ? "; Because: " + why : "") + "; " + this);
}
}
public synchronized void requestSystemCompaction(HRegion region, String why) throws IOException {
requestCompactionInternal(region, why, Store.NO_PRIORITY, false,
CompactionLifeCycleTracker.DUMMY, null);
}
public synchronized void requestSystemCompaction(HRegion region, HStore store, String why)
throws IOException {
requestCompactionInternal(region, store, why, Store.NO_PRIORITY, false,
CompactionLifeCycleTracker.DUMMY, null);
}
private Optional selectCompaction(HRegion region, HStore store, int priority,
CompactionLifeCycleTracker tracker, User user) throws IOException {
Optional compaction = store.requestCompaction(priority, tracker, user);
if (!compaction.isPresent() && LOG.isDebugEnabled() && region.getRegionInfo() != null) {
LOG.debug("Not compacting " + region.getRegionInfo().getRegionNameAsString() +
" because compaction request was cancelled");
}
return compaction;
}
/**
* Only interrupt once it's done with a run through the work loop.
*/
void interruptIfNecessary() {
splits.shutdown();
longCompactions.shutdown();
shortCompactions.shutdown();
}
private void waitFor(ThreadPoolExecutor t, String name) {
boolean done = false;
while (!done) {
try {
done = t.awaitTermination(60, TimeUnit.SECONDS);
LOG.info("Waiting for " + name + " to finish...");
if (!done) {
t.shutdownNow();
}
} catch (InterruptedException ie) {
LOG.warn("Interrupted waiting for " + name + " to finish...");
}
}
}
void join() {
waitFor(splits, "Split Thread");
waitFor(longCompactions, "Large Compaction Thread");
waitFor(shortCompactions, "Small Compaction Thread");
}
/**
* Returns the current size of the queue containing regions that are
* processed.
*
* @return The current size of the regions queue.
*/
public int getCompactionQueueSize() {
return longCompactions.getQueue().size() + shortCompactions.getQueue().size();
}
public int getLargeCompactionQueueSize() {
return longCompactions.getQueue().size();
}
public int getSmallCompactionQueueSize() {
return shortCompactions.getQueue().size();
}
public int getSplitQueueSize() {
return splits.getQueue().size();
}
private boolean shouldSplitRegion() {
if(server.getNumberOfOnlineRegions() > 0.9*regionSplitLimit) {
LOG.warn("Total number of regions is approaching the upper limit " + regionSplitLimit + ". "
+ "Please consider taking a look at http://hbase.apache.org/book.html#ops.regionmgt");
}
return (regionSplitLimit > server.getNumberOfOnlineRegions());
}
/**
* @return the regionSplitLimit
*/
public int getRegionSplitLimit() {
return this.regionSplitLimit;
}
private static final Comparator COMPARATOR =
new Comparator() {
private int compare(CompactionRequest r1, CompactionRequest r2) {
if (r1 == r2) {
return 0; //they are the same request
}
// less first
int cmp = Integer.compare(r1.getPriority(), r2.getPriority());
if (cmp != 0) {
return cmp;
}
cmp = Long.compare(r1.getSelectionNanoTime(), r2.getSelectionNanoTime());
if (cmp != 0) {
return cmp;
}
// break the tie based on hash code
return System.identityHashCode(r1) - System.identityHashCode(r2);
}
@Override
public int compare(Runnable r1, Runnable r2) {
// CompactionRunner first
if (r1 instanceof CompactionRunner) {
if (!(r2 instanceof CompactionRunner)) {
return -1;
}
} else {
if (r2 instanceof CompactionRunner) {
return 1;
} else {
// break the tie based on hash code
return System.identityHashCode(r1) - System.identityHashCode(r2);
}
}
CompactionRunner o1 = (CompactionRunner) r1;
CompactionRunner o2 = (CompactionRunner) r2;
// less first
int cmp = Integer.compare(o1.queuedPriority, o2.queuedPriority);
if (cmp != 0) {
return cmp;
}
Optional c1 = o1.compaction;
Optional c2 = o2.compaction;
if (c1.isPresent()) {
return c2.isPresent() ? compare(c1.get().getRequest(), c2.get().getRequest()) : -1;
} else {
return c2.isPresent() ? 1 : 0;
}
}
};
private final class CompactionRunner implements Runnable {
private final HStore store;
private final HRegion region;
private final Optional compaction;
private int queuedPriority;
private ThreadPoolExecutor parent;
private User user;
private long time;
public CompactionRunner(HStore store, HRegion region, Optional compaction,
ThreadPoolExecutor parent, User user) {
super();
this.store = store;
this.region = region;
this.compaction = compaction;
this.queuedPriority = compaction.isPresent() ? compaction.get().getRequest().getPriority()
: store.getCompactPriority();
this.parent = parent;
this.user = user;
this.time = System.currentTimeMillis();
}
@Override
public String toString() {
return compaction.map(c -> "Request = " + c.getRequest())
.orElse("regionName = " + region.toString() + ", storeName = " + store.toString() +
", priority = " + queuedPriority + ", time = " + time);
}
private void doCompaction(User user) {
CompactionContext c;
// Common case - system compaction without a file selection. Select now.
if (!compaction.isPresent()) {
int oldPriority = this.queuedPriority;
this.queuedPriority = this.store.getCompactPriority();
if (this.queuedPriority > oldPriority) {
// Store priority decreased while we were in queue (due to some other compaction?),
// requeue with new priority to avoid blocking potential higher priorities.
this.parent.execute(this);
return;
}
Optional selected;
try {
selected = selectCompaction(this.region, this.store, queuedPriority,
CompactionLifeCycleTracker.DUMMY, user);
} catch (IOException ex) {
LOG.error("Compaction selection failed " + this, ex);
server.checkFileSystem();
region.decrementCompactionsQueuedCount();
return;
}
if (!selected.isPresent()) {
region.decrementCompactionsQueuedCount();
return; // nothing to do
}
c = selected.get();
assert c.hasSelection();
// Now see if we are in correct pool for the size; if not, go to the correct one.
// We might end up waiting for a while, so cancel the selection.
ThreadPoolExecutor pool =
store.throttleCompaction(c.getRequest().getSize()) ? longCompactions : shortCompactions;
// Long compaction pool can process small job
// Short compaction pool should not process large job
if (this.parent == shortCompactions && pool == longCompactions) {
this.store.cancelRequestedCompaction(c);
this.parent = pool;
this.parent.execute(this);
return;
}
} else {
c = compaction.get();
}
// Finally we can compact something.
assert c != null;
c.getRequest().getTracker().beforeExecute(store);
try {
// Note: please don't put single-compaction logic here;
// put it into region/store/etc. This is CST logic.
long start = EnvironmentEdgeManager.currentTime();
boolean completed =
region.compact(c, store, compactionThroughputController, user);
long now = EnvironmentEdgeManager.currentTime();
LOG.info(((completed) ? "Completed" : "Aborted") + " compaction: " +
this + "; duration=" + StringUtils.formatTimeDiff(now, start));
if (completed) {
// degenerate case: blocked regions require recursive enqueues
if (store.getCompactPriority() <= 0) {
requestSystemCompaction(region, store, "Recursive enqueue");
} else {
// see if the compaction has caused us to exceed max region size
requestSplit(region);
}
}
} catch (IOException ex) {
IOException remoteEx =
ex instanceof RemoteException ? ((RemoteException) ex).unwrapRemoteException() : ex;
LOG.error("Compaction failed " + this, remoteEx);
if (remoteEx != ex) {
LOG.info("Compaction failed at original callstack: " + formatStackTrace(ex));
}
region.reportCompactionRequestFailure();
server.checkFileSystem();
} catch (Exception ex) {
LOG.error("Compaction failed " + this, ex);
region.reportCompactionRequestFailure();
server.checkFileSystem();
} finally {
c.getRequest().getTracker().afterExecute(store);
region.decrementCompactionsQueuedCount();
LOG.debug("CompactSplitThread Status: " + CompactSplit.this);
}
}
@Override
public void run() {
Preconditions.checkNotNull(server);
if (server.isStopped()
|| (region.getTableDescriptor() != null && !region.getTableDescriptor().isCompactionEnabled())) {
region.decrementCompactionsQueuedCount();
return;
}
doCompaction(user);
}
private String formatStackTrace(Exception ex) {
StringWriter sw = new StringWriter();
PrintWriter pw = new PrintWriter(sw);
ex.printStackTrace(pw);
pw.flush();
return sw.toString();
}
}
/**
* Cleanup class to use when rejecting a compaction request from the queue.
*/
private static class Rejection implements RejectedExecutionHandler {
@Override
public void rejectedExecution(Runnable runnable, ThreadPoolExecutor pool) {
if (runnable instanceof CompactionRunner) {
CompactionRunner runner = (CompactionRunner) runnable;
LOG.debug("Compaction Rejected: " + runner);
runner.compaction.ifPresent(c -> runner.store.cancelRequestedCompaction(c));
}
}
}
/**
* {@inheritDoc}
*/
@Override
public void onConfigurationChange(Configuration newConf) {
// Check if number of large / small compaction threads has changed, and then
// adjust the core pool size of the thread pools, by using the
// setCorePoolSize() method. According to the javadocs, it is safe to
// change the core pool size on-the-fly. We need to reset the maximum
// pool size, as well.
int largeThreads = Math.max(1, newConf.getInt(
LARGE_COMPACTION_THREADS,
LARGE_COMPACTION_THREADS_DEFAULT));
if (this.longCompactions.getCorePoolSize() != largeThreads) {
LOG.info("Changing the value of " + LARGE_COMPACTION_THREADS +
" from " + this.longCompactions.getCorePoolSize() + " to " +
largeThreads);
if(this.longCompactions.getCorePoolSize() < largeThreads) {
this.longCompactions.setMaximumPoolSize(largeThreads);
this.longCompactions.setCorePoolSize(largeThreads);
} else {
this.longCompactions.setCorePoolSize(largeThreads);
this.longCompactions.setMaximumPoolSize(largeThreads);
}
}
int smallThreads = newConf.getInt(SMALL_COMPACTION_THREADS,
SMALL_COMPACTION_THREADS_DEFAULT);
if (this.shortCompactions.getCorePoolSize() != smallThreads) {
LOG.info("Changing the value of " + SMALL_COMPACTION_THREADS +
" from " + this.shortCompactions.getCorePoolSize() + " to " +
smallThreads);
if(this.shortCompactions.getCorePoolSize() < smallThreads) {
this.shortCompactions.setMaximumPoolSize(smallThreads);
this.shortCompactions.setCorePoolSize(smallThreads);
} else {
this.shortCompactions.setCorePoolSize(smallThreads);
this.shortCompactions.setMaximumPoolSize(smallThreads);
}
}
int splitThreads = newConf.getInt(SPLIT_THREADS,
SPLIT_THREADS_DEFAULT);
if (this.splits.getCorePoolSize() != splitThreads) {
LOG.info("Changing the value of " + SPLIT_THREADS +
" from " + this.splits.getCorePoolSize() + " to " +
splitThreads);
if(this.splits.getCorePoolSize() < splitThreads) {
this.splits.setMaximumPoolSize(splitThreads);
this.splits.setCorePoolSize(splitThreads);
} else {
this.splits.setCorePoolSize(splitThreads);
this.splits.setMaximumPoolSize(splitThreads);
}
}
ThroughputController old = this.compactionThroughputController;
if (old != null) {
old.stop("configuration change");
}
this.compactionThroughputController =
CompactionThroughputControllerFactory.create(server, newConf);
// We change this atomically here instead of reloading the config in order that upstream
// would be the only one with the flexibility to reload the config.
this.conf.reloadConfiguration();
}
protected int getSmallCompactionThreadNum() {
return this.shortCompactions.getCorePoolSize();
}
protected int getLargeCompactionThreadNum() {
return this.longCompactions.getCorePoolSize();
}
protected int getSplitThreadNum() {
return this.splits.getCorePoolSize();
}
/**
* {@inheritDoc}
*/
@Override
public void registerChildren(ConfigurationManager manager) {
// No children to register.
}
/**
* {@inheritDoc}
*/
@Override
public void deregisterChildren(ConfigurationManager manager) {
// No children to register
}
@VisibleForTesting
public ThroughputController getCompactionThroughputController() {
return compactionThroughputController;
}
@VisibleForTesting
/**
* Shutdown the long compaction thread pool.
* Should only be used in unit test to prevent long compaction thread pool from stealing job
* from short compaction queue
*/
void shutdownLongCompactions(){
this.longCompactions.shutdown();
}
public void clearLongCompactionsQueue() {
longCompactions.getQueue().clear();
}
public void clearShortCompactionsQueue() {
shortCompactions.getQueue().clear();
}
}