系列
- RocketMq broker 配置文件
- RocketMq broker 启动流程
- RocketMq broker CommitLog介绍
- RocketMq broker consumeQueue介绍
- RocketMq broker 重试和死信队列
- RocketMq broker 延迟消息
- RocketMq IndexService介绍
- RocketMq 读写分离机制
- RocketMq Client管理
- RocketMq Broker线程模型及快速失败机制
- RocketMq broker过期文件删除
开篇
这个系列的主要目的是介绍RocketMq broker的原理和用法，在这个系列当中会介绍 broker 配置文件、broker 启动流程、broker延迟消息、broker消息存储。
这篇文章主要介绍broker 消息存储，主要介绍broker在处理消息发送请求的整个流程中核心的consumeQueue的作用。
ConsumeQueue：消息消费队列，引入的目的主要是提高消息消费的性能，由于RocketMQ是基于主题topic的订阅模式，消息消费是针对主题进行的，如果要遍历commitlog文件中根据topic检索消息是非常低效的。Consumer即可根据ConsumeQueue来查找待消费的消息。其中，ConsumeQueue（逻辑消费队列）作为消费消息的索引，保存了指定Topic下的队列消息在CommitLog中的起始物理偏移量offset，消息大小size和消息Tag的HashCode值。
consumequeue文件可以看成是基于topic的commitlog索引文件，故consumequeue文件夹的组织方式如下：topic/queue/file三层组织结构，具体存储路径为：$HOME/store/consumequeue/{topic}/{queueId}/{fileName}。同样consumequeue文件采取定长设计，每一个条目共20个字节，分别为8字节的commitlog物理偏移量、4字节的消息长度、8字节tag hashcode，单个文件由30W个条目组成，可以像数组一样随机访问每一个条目，每个ConsumeQueue文件大小约5.72M。
DefaultMessageStore#start
public class DefaultMessageStore implements MessageStore {
public void start() throws Exception {
lock = lockFile.getChannel().tryLock(0, 1, false);
if (lock == null || lock.isShared() || !lock.isValid()) {
throw new RuntimeException("Lock failed,MQ already started");
}
lockFile.getChannel().write(ByteBuffer.wrap("lock".getBytes()));
lockFile.getChannel().force(true);
{
/**
* 1. Make sure the fast-forward messages to be truncated during the recovering according to the max physical offset of the commitlog;
* 2. DLedger committedPos may be missing, so the maxPhysicalPosInLogicQueue maybe bigger that maxOffset returned by DLedgerCommitLog, just let it go;
* 3. Calculate the reput offset according to the consume queue;
* 4. Make sure the fall-behind messages to be dispatched before starting the commitlog, especially when the broker role are automatically changed.
*/
long maxPhysicalPosInLogicQueue = commitLog.getMinOffset();
for (ConcurrentMap maps : this.consumeQueueTable.values()) {
for (ConsumeQueue logic : maps.values()) {
if (logic.getMaxPhysicOffset() > maxPhysicalPosInLogicQueue) {
maxPhysicalPosInLogicQueue = logic.getMaxPhysicOffset();
}
}
}
if (maxPhysicalPosInLogicQueue < 0) {
maxPhysicalPosInLogicQueue = 0;
}
if (maxPhysicalPosInLogicQueue < this.commitLog.getMinOffset()) {
maxPhysicalPosInLogicQueue = this.commitLog.getMinOffset();
/**
* This happens in following conditions:
* 1. If someone removes all the consumequeue files or the disk get damaged.
* 2. Launch a new broker, and copy the commitlog from other brokers.
*
* All the conditions has the same in common that the maxPhysicalPosInLogicQueue should be 0.
* If the maxPhysicalPosInLogicQueue is gt 0, there maybe something wrong.
*/
log.warn("[TooSmallCqOffset] maxPhysicalPosInLogicQueue={} clMinOffset={}", maxPhysicalPosInLogicQueue, this.commitLog.getMinOffset());
}
log.info("[SetReputOffset] maxPhysicalPosInLogicQueue={} clMinOffset={} clMaxOffset={} clConfirmedOffset={}",
maxPhysicalPosInLogicQueue, this.commitLog.getMinOffset(), this.commitLog.getMaxOffset(), this.commitLog.getConfirmOffset());
this.reputMessageService.setReputFromOffset(maxPhysicalPosInLogicQueue);
// 启动reputMessageService
this.reputMessageService.start();
/**
* 1. Finish dispatching the messages fall behind, then to start other services.
* 2. DLedger committedPos may be missing, so here just require dispatchBehindBytes <= 0
*/
while (true) {
if (dispatchBehindBytes() <= 0) {
break;
}
Thread.sleep(1000);
log.info("Try to finish doing reput the messages fall behind during the starting, reputOffset={} maxOffset={} behind={}", this.reputMessageService.getReputFromOffset(), this.getMaxPhyOffset(), this.dispatchBehindBytes());
}
this.recoverTopicQueueTable();
}
if (!messageStoreConfig.isEnableDLegerCommitLog()) {
this.haService.start();
this.handleScheduleMessageService(messageStoreConfig.getBrokerRole());
}
// 2、启动flushConsumeQueueService
this.flushConsumeQueueService.start();
this.commitLog.start();
this.storeStatsService.start();
this.createTempFile();
this.addScheduleTask();
this.shutdown = false;
}
}
- reputMessageService.start()负责启动reputMessageService，负责读取commitLog信息并添加到consumeQueue当中。
ReputMessageService
public class DefaultMessageStore implements MessageStore {
class ReputMessageService extends ServiceThread {
private volatile long reputFromOffset = 0;
public long getReputFromOffset() {
return reputFromOffset;
}
public void setReputFromOffset(long reputFromOffset) {
this.reputFromOffset = reputFromOffset;
}
private void doReput() {
if (this.reputFromOffset < DefaultMessageStore.this.commitLog.getMinOffset()) {
log.warn("The reputFromOffset={} is smaller than minPyOffset={}, this usually indicate that the dispatch behind too much and the commitlog has expired.",
this.reputFromOffset, DefaultMessageStore.this.commitLog.getMinOffset());
this.reputFromOffset = DefaultMessageStore.this.commitLog.getMinOffset();
}
for (boolean doNext = true; this.isCommitLogAvailable() && doNext; ) {
if (DefaultMessageStore.this.getMessageStoreConfig().isDuplicationEnable()
&& this.reputFromOffset >= DefaultMessageStore.this.getConfirmOffset()) {
break;
}
// 1、获取待put的数据对象SelectMappedBufferResult
SelectMappedBufferResult result = DefaultMessageStore.this.commitLog.getData(reputFromOffset);
// 2、解析获取的数据结果
if (result != null) {
try {
this.reputFromOffset = result.getStartOffset();
for (int readSize = 0; readSize < result.getSize() && doNext; ) {
DispatchRequest dispatchRequest =
DefaultMessageStore.this.commitLog.checkMessageAndReturnSize(result.getByteBuffer(), false, false);
int size = dispatchRequest.getBufferSize() == -1 ? dispatchRequest.getMsgSize() : dispatchRequest.getBufferSize();
if (dispatchRequest.isSuccess()) {
if (size > 0) {
// 3、保存索引数据到consumeQueue
DefaultMessageStore.this.doDispatch(dispatchRequest);
if (BrokerRole.SLAVE != DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole()
&& DefaultMessageStore.this.brokerConfig.isLongPollingEnable()) {
DefaultMessageStore.this.messageArrivingListener.arriving(dispatchRequest.getTopic(),
dispatchRequest.getQueueId(), dispatchRequest.getConsumeQueueOffset() + 1,
dispatchRequest.getTagsCode(), dispatchRequest.getStoreTimestamp(),
dispatchRequest.getBitMap(), dispatchRequest.getPropertiesMap());
}
this.reputFromOffset += size;
readSize += size;
if (DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole() == BrokerRole.SLAVE) {
DefaultMessageStore.this.storeStatsService
.getSinglePutMessageTopicTimesTotal(dispatchRequest.getTopic()).incrementAndGet();
DefaultMessageStore.this.storeStatsService
.getSinglePutMessageTopicSizeTotal(dispatchRequest.getTopic())
.addAndGet(dispatchRequest.getMsgSize());
}
} else if (size == 0) {
this.reputFromOffset = DefaultMessageStore.this.commitLog.rollNextFile(this.reputFromOffset);
readSize = result.getSize();
}
} else if (!dispatchRequest.isSuccess()) {
if (size > 0) {
log.error("[BUG]read total count not equals msg total size. reputFromOffset={}", reputFromOffset);
this.reputFromOffset += size;
} else {
doNext = false;
// If user open the dledger pattern or the broker is master node,
// it will not ignore the exception and fix the reputFromOffset variable
if (DefaultMessageStore.this.getMessageStoreConfig().isEnableDLegerCommitLog() ||
DefaultMessageStore.this.brokerConfig.getBrokerId() == MixAll.MASTER_ID) {
log.error("[BUG]dispatch message to consume queue error, COMMITLOG OFFSET: {}",
this.reputFromOffset);
this.reputFromOffset += result.getSize() - readSize;
}
}
}
}
} finally {
result.release();
}
} else {
doNext = false;
}
}
}
@Override
public void run() {
DefaultMessageStore.log.info(this.getServiceName() + " service started");
while (!this.isStopped()) {
try {
Thread.sleep(1);
this.doReput();
} catch (Exception e) {
DefaultMessageStore.log.warn(this.getServiceName() + " service has exception. ", e);
}
}
DefaultMessageStore.log.info(this.getServiceName() + " service end");
}
@Override
public String getServiceName() {
return ReputMessageService.class.getSimpleName();
}
}
}
- ReputMessageService#run内部执行doReput方法。
- doReput方法负责从commitLog获取消息并将位移等相关信息保存到consumeQueue。
- 每次从reputFromOffset位置开始读取commitLog。
CommitLog数据读取
public class CommitLog {
protected final AtomicInteger wrotePosition = new AtomicInteger(0);
public SelectMappedBufferResult getData(final long offset) {
return this.getData(offset, offset == 0);
}
public SelectMappedBufferResult getData(final long offset, final boolean returnFirstOnNotFound) {
// 获取固定的mappedFileSize
int mappedFileSize = this.defaultMessageStore.getMessageStoreConfig().getMappedFileSizeCommitLog();
// 获取MappedFile
MappedFile mappedFile = this.mappedFileQueue.findMappedFileByOffset(offset, returnFirstOnNotFound);
if (mappedFile != null) {
// 从MappedFile的pos位置开始查找
int pos = (int) (offset % mappedFileSize);
SelectMappedBufferResult result = mappedFile.selectMappedBuffer(pos);
return result;
}
return null;
}
}
public class MappedFile extends ReferenceResource {
protected final AtomicInteger wrotePosition = new AtomicInteger(0);
public int getReadPosition() {
return this.writeBuffer == null ? this.wrotePosition.get() : this.committedPosition.get();
}
public SelectMappedBufferResult selectMappedBuffer(int pos) {
// 获取能够读取的数据的最大位移
int readPosition = getReadPosition();
// 获取从pos 到 readPosition之间的数据
if (pos < readPosition && pos >= 0) {
if (this.hold()) {
ByteBuffer byteBuffer = this.mappedByteBuffer.slice();
byteBuffer.position(pos);
int size = readPosition - pos;
ByteBuffer byteBufferNew = byteBuffer.slice();
byteBufferNew.limit(size);
return new SelectMappedBufferResult(this.fileFromOffset + pos, byteBufferNew, size, this);
}
}
return null;
}
}
public class SelectMappedBufferResult {
private final long startOffset;
private final ByteBuffer byteBuffer;
private int size;
private MappedFile mappedFile;
public SelectMappedBufferResult(long startOffset, ByteBuffer byteBuffer, int size, MappedFile mappedFile) {
this.startOffset = startOffset;
this.byteBuffer = byteBuffer;
this.size = size;
this.mappedFile = mappedFile;
}
}
- CommitLog#getData根据起始offset的位移去mappedFileQueue确定mappedFile文件。
- MappedFile.selectMappedBuffer获取目前数据写入的最大位移，根据起始offset和结束位移。
- MappedFile的目前数据写入的最大位移wrotePosition会在消息写入后变更，每次数据添加后同步更新wrotePosition的值。
- getData返回SelectMappedBufferResult对象，包括byteBuffer（数据缓存）、startOffset（起始位移）、size（数据大小）、mappedFile（对应的mappedFile）。
consumeQueue的写入
public class DefaultMessageStore implements MessageStore {
class ReputMessageService extends ServiceThread {
private void doReput() {
if (this.reputFromOffset < DefaultMessageStore.this.commitLog.getMinOffset()) {
log.warn("The reputFromOffset={} is smaller than minPyOffset={}, this usually indicate that the dispatch behind too much and the commitlog has expired.",
this.reputFromOffset, DefaultMessageStore.this.commitLog.getMinOffset());
this.reputFromOffset = DefaultMessageStore.this.commitLog.getMinOffset();
}
for (boolean doNext = true; this.isCommitLogAvailable() && doNext; ) {
if (DefaultMessageStore.this.getMessageStoreConfig().isDuplicationEnable()
&& this.reputFromOffset >= DefaultMessageStore.this.getConfirmOffset()) {
break;
}
SelectMappedBufferResult result = DefaultMessageStore.this.commitLog.getData(reputFromOffset);
// 解析获取的数据结果
if (result != null) {
try {
this.reputFromOffset = result.getStartOffset();
for (int readSize = 0; readSize < result.getSize() && doNext; ) {
// 负责解析SelectMappedBufferResult生成DispatchRequest
DispatchRequest dispatchRequest =
DefaultMessageStore.this.commitLog.checkMessageAndReturnSize(result.getByteBuffer(), false, false);
int size = dispatchRequest.getBufferSize() == -1 ? dispatchRequest.getMsgSize() : dispatchRequest.getBufferSize();
if (dispatchRequest.isSuccess()) {
if (size > 0) {
// 保存dispatchRequest
DefaultMessageStore.this.doDispatch(dispatchRequest);
if (BrokerRole.SLAVE != DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole()
&& DefaultMessageStore.this.brokerConfig.isLongPollingEnable()) {
DefaultMessageStore.this.messageArrivingListener.arriving(dispatchRequest.getTopic(),
dispatchRequest.getQueueId(), dispatchRequest.getConsumeQueueOffset() + 1,
dispatchRequest.getTagsCode(), dispatchRequest.getStoreTimestamp(),
dispatchRequest.getBitMap(), dispatchRequest.getPropertiesMap());
}
this.reputFromOffset += size;
readSize += size;
if (DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole() == BrokerRole.SLAVE) {
DefaultMessageStore.this.storeStatsService
.getSinglePutMessageTopicTimesTotal(dispatchRequest.getTopic()).incrementAndGet();
DefaultMessageStore.this.storeStatsService
.getSinglePutMessageTopicSizeTotal(dispatchRequest.getTopic())
.addAndGet(dispatchRequest.getMsgSize());
}
} else if (size == 0) {
this.reputFromOffset = DefaultMessageStore.this.commitLog.rollNextFile(this.reputFromOffset);
readSize = result.getSize();
}
} else if (!dispatchRequest.isSuccess()) {
if (size > 0) {
log.error("[BUG]read total count not equals msg total size. reputFromOffset={}", reputFromOffset);
this.reputFromOffset += size;
} else {
doNext = false;
// If user open the dledger pattern or the broker is master node,
// it will not ignore the exception and fix the reputFromOffset variable
if (DefaultMessageStore.this.getMessageStoreConfig().isEnableDLegerCommitLog() ||
DefaultMessageStore.this.brokerConfig.getBrokerId() == MixAll.MASTER_ID) {
log.error("[BUG]dispatch message to consume queue error, COMMITLOG OFFSET: {}",
this.reputFromOffset);
this.reputFromOffset += result.getSize() - readSize;
}
}
}
}
} finally {
result.release();
}
} else {
doNext = false;
}
}
}
}
}
- commitLog.checkMessageAndReturnSize负责解析报文。
- DefaultMessageStore.this.doDispatch(dispatchRequest)负责保存解析结果。
commitLog解析
public class CommitLog {
public DispatchRequest checkMessageAndReturnSize(java.nio.ByteBuffer byteBuffer, final boolean checkCRC,
final boolean readBody) {
try {
// 1 TOTAL SIZE
int totalSize = byteBuffer.getInt();
// 2 MAGIC CODE
int magicCode = byteBuffer.getInt();
switch (magicCode) {
case MESSAGE_MAGIC_CODE:
break;
case BLANK_MAGIC_CODE:
return new DispatchRequest(0, true /* success */);
default:
log.warn("found a illegal magic code 0x" + Integer.toHexString(magicCode));
return new DispatchRequest(-1, false /* success */);
}
byte[] bytesContent = new byte[totalSize];
int bodyCRC = byteBuffer.getInt();
int queueId = byteBuffer.getInt();
int flag = byteBuffer.getInt();
long queueOffset = byteBuffer.getLong();
long physicOffset = byteBuffer.getLong();
int sysFlag = byteBuffer.getInt();
long bornTimeStamp = byteBuffer.getLong();
ByteBuffer byteBuffer1;
if ((sysFlag & MessageSysFlag.BORNHOST_V6_FLAG) == 0) {
byteBuffer1 = byteBuffer.get(bytesContent, 0, 4 + 4);
} else {
byteBuffer1 = byteBuffer.get(bytesContent, 0, 16 + 4);
}
long storeTimestamp = byteBuffer.getLong();
ByteBuffer byteBuffer2;
if ((sysFlag & MessageSysFlag.STOREHOSTADDRESS_V6_FLAG) == 0) {
byteBuffer2 = byteBuffer.get(bytesContent, 0, 4 + 4);
} else {
byteBuffer2 = byteBuffer.get(bytesContent, 0, 16 + 4);
}
int reconsumeTimes = byteBuffer.getInt();
long preparedTransactionOffset = byteBuffer.getLong();
int bodyLen = byteBuffer.getInt();
if (bodyLen > 0) {
if (readBody) {
byteBuffer.get(bytesContent, 0, bodyLen);
if (checkCRC) {
int crc = UtilAll.crc32(bytesContent, 0, bodyLen);
if (crc != bodyCRC) {
log.warn("CRC check failed. bodyCRC={}, currentCRC={}", crc, bodyCRC);
return new DispatchRequest(-1, false/* success */);
}
}
} else {
byteBuffer.position(byteBuffer.position() + bodyLen);
}
}
byte topicLen = byteBuffer.get();
byteBuffer.get(bytesContent, 0, topicLen);
String topic = new String(bytesContent, 0, topicLen, MessageDecoder.CHARSET_UTF8);
long tagsCode = 0;
String keys = "";
String uniqKey = null;
short propertiesLength = byteBuffer.getShort();
Map propertiesMap = null;
if (propertiesLength > 0) {
byteBuffer.get(bytesContent, 0, propertiesLength);
String properties = new String(bytesContent, 0, propertiesLength, MessageDecoder.CHARSET_UTF8);
propertiesMap = MessageDecoder.string2messageProperties(properties);
keys = propertiesMap.get(MessageConst.PROPERTY_KEYS);
uniqKey = propertiesMap.get(MessageConst.PROPERTY_UNIQ_CLIENT_MESSAGE_ID_KEYIDX);
String tags = propertiesMap.get(MessageConst.PROPERTY_TAGS);
if (tags != null && tags.length() > 0) {
tagsCode = MessageExtBrokerInner.tagsString2tagsCode(MessageExt.parseTopicFilterType(sysFlag), tags);
}
// Timing message processing
{
String t = propertiesMap.get(MessageConst.PROPERTY_DELAY_TIME_LEVEL);
if (ScheduleMessageService.SCHEDULE_TOPIC.equals(topic) && t != null) {
int delayLevel = Integer.parseInt(t);
if (delayLevel > this.defaultMessageStore.getScheduleMessageService().getMaxDelayLevel()) {
delayLevel = this.defaultMessageStore.getScheduleMessageService().getMaxDelayLevel();
}
if (delayLevel > 0) {
tagsCode = this.defaultMessageStore.getScheduleMessageService().computeDeliverTimestamp(delayLevel,
storeTimestamp);
}
}
}
}
int readLength = calMsgLength(sysFlag, bodyLen, topicLen, propertiesLength);
if (totalSize != readLength) {
doNothingForDeadCode(reconsumeTimes);
doNothingForDeadCode(flag);
doNothingForDeadCode(bornTimeStamp);
doNothingForDeadCode(byteBuffer1);
doNothingForDeadCode(byteBuffer2);
log.error(
"[BUG]read total count not equals msg total size. totalSize={}, readTotalCount={}, bodyLen={}, topicLen={}, propertiesLength={}",
totalSize, readLength, bodyLen, topicLen, propertiesLength);
return new DispatchRequest(totalSize, false/* success */);
}
return new DispatchRequest(
topic,
queueId,
physicOffset,
totalSize,
tagsCode,
storeTimestamp,
queueOffset,
keys,
uniqKey,
sysFlag,
preparedTransactionOffset,
propertiesMap
);
} catch (Exception e) {
}
return new DispatchRequest(-1, false /* success */);
}
}
public class DispatchRequest {
private final String topic;
private final int queueId;
private final long commitLogOffset;
private int msgSize;
private final long tagsCode;
private final long storeTimestamp;
private final long consumeQueueOffset;
private final String keys;
private final boolean success;
private final String uniqKey;
private final int sysFlag;
private final long preparedTransactionOffset;
private final Map propertiesMap;
private byte[] bitMap;
private int bufferSize = -1;//the buffer size maybe larger than the msg size if the message is wrapped by something
public DispatchRequest(
final String topic,
final int queueId,
final long commitLogOffset,
final int msgSize,
final long tagsCode,
final long storeTimestamp,
final long consumeQueueOffset,
final String keys,
final String uniqKey,
final int sysFlag,
final long preparedTransactionOffset,
final Map propertiesMap
) {
this.topic = topic;
this.queueId = queueId;
this.commitLogOffset = commitLogOffset;
this.msgSize = msgSize;
this.tagsCode = tagsCode;
this.storeTimestamp = storeTimestamp;
this.consumeQueueOffset = consumeQueueOffset;
this.keys = keys;
this.uniqKey = uniqKey;
this.sysFlag = sysFlag;
this.preparedTransactionOffset = preparedTransactionOffset;
this.success = true;
this.propertiesMap = propertiesMap;
}
}
- checkMessageAndReturnSize实现commitLog的解析，按照commitLog的报文格式进行解析，解析的生成的DispatchRequest对象。
- commitLogOffset保存commitLog实际的偏移。
- consumeQueueOffset保存在consumeQueue的位移。
- consumeQueue这个关键字段是保存在CommitLog类的topicQueueTable当中，
HashMap/, Long/ offset */> topicQueueTable，相当于指定topic下的指定queueId下已保存的消息各数。
doDispatch分发任务
public class DefaultMessageStore implements MessageStore {
private final LinkedList dispatcherList;
public DefaultMessageStore(final MessageStoreConfig messageStoreConfig, final BrokerStatsManager brokerStatsManager,
final MessageArrivingListener messageArrivingListener, final BrokerConfig brokerConfig) throws IOException {
this.allocateMappedFileService.start();
this.indexService.start();
this.dispatcherList = new LinkedList<>();
this.dispatcherList.addLast(new CommitLogDispatcherBuildConsumeQueue());
this.dispatcherList.addLast(new CommitLogDispatcherBuildIndex());
File file = new File(StorePathConfigHelper.getLockFile(messageStoreConfig.getStorePathRootDir()));
MappedFile.ensureDirOK(file.getParent());
lockFile = new RandomAccessFile(file, "rw");
}
public void doDispatch(DispatchRequest req) {
for (CommitLogDispatcher dispatcher : this.dispatcherList) {
dispatcher.dispatch(req);
}
}
}
- doDispatch负责执行分发任务。
- CommitLogDispatcherBuildConsumeQueue负责生成consumeQueue。
- CommitLogDispatcherBuildIndex负责生成Index。
CommitLogDispatcherBuildConsumeQueue
public class DefaultMessageStore implements MessageStore {
class CommitLogDispatcherBuildConsumeQueue implements CommitLogDispatcher {
@Override
public void dispatch(DispatchRequest request) {
final int tranType = MessageSysFlag.getTransactionValue(request.getSysFlag());
switch (tranType) {
case MessageSysFlag.TRANSACTION_NOT_TYPE:
case MessageSysFlag.TRANSACTION_COMMIT_TYPE:
DefaultMessageStore.this.putMessagePositionInfo(request);
break;
case MessageSysFlag.TRANSACTION_PREPARED_TYPE:
case MessageSysFlag.TRANSACTION_ROLLBACK_TYPE:
break;
}
}
}
}
- dispatch执行DefaultMessageStore#putMessagePositionInfo。
public class DefaultMessageStore implements MessageStore {
private final ConcurrentMap> consumeQueueTable;
public void putMessagePositionInfo(DispatchRequest dispatchRequest) {
ConsumeQueue cq = this.findConsumeQueue(dispatchRequest.getTopic(), dispatchRequest.getQueueId());
cq.putMessagePositionInfoWrapper(dispatchRequest);
}
public ConsumeQueue findConsumeQueue(String topic, int queueId) {
ConcurrentMap map = consumeQueueTable.get(topic);
if (null == map) {
ConcurrentMap newMap = new ConcurrentHashMap(128);
ConcurrentMap oldMap = consumeQueueTable.putIfAbsent(topic, newMap);
if (oldMap != null) {
map = oldMap;
} else {
map = newMap;
}
}
ConsumeQueue logic = map.get(queueId);
if (null == logic) {
ConsumeQueue newLogic = new ConsumeQueue(
topic,
queueId,
StorePathConfigHelper.getStorePathConsumeQueue(this.messageStoreConfig.getStorePathRootDir()),
this.getMessageStoreConfig().getMappedFileSizeConsumeQueue(),
this);
ConsumeQueue oldLogic = map.putIfAbsent(queueId, newLogic);
if (oldLogic != null) {
logic = oldLogic;
} else {
logic = newLogic;
}
}
return logic;
}
}
- putMessagePositionInfo核心逻辑是寻找ConsumeQueue对象。
- consumeQueueTable保存的以topic+queueId保存ConsumeQueue。
- ConsumeQueue保存mappedFileQueue，mappedFileQueue保存mappedFiles。
- ConsumeQueue本质上还是通过mappedFile来保存的。
public class ConsumeQueue {
public static final int CQ_STORE_UNIT_SIZE = 20;
public void putMessagePositionInfoWrapper(DispatchRequest request) {
final int maxRetries = 30;
boolean canWrite = this.defaultMessageStore.getRunningFlags().isCQWriteable();
for (int i = 0; i < maxRetries && canWrite; i++) {
long tagsCode = request.getTagsCode();
if (isExtWriteEnable()) {
ConsumeQueueExt.CqExtUnit cqExtUnit = new ConsumeQueueExt.CqExtUnit();
cqExtUnit.setFilterBitMap(request.getBitMap());
cqExtUnit.setMsgStoreTime(request.getStoreTimestamp());
cqExtUnit.setTagsCode(request.getTagsCode());
long extAddr = this.consumeQueueExt.put(cqExtUnit);
if (isExtAddr(extAddr)) {
tagsCode = extAddr;
} else {
log.warn("Save consume queue extend fail, So just save tagsCode! {}, topic:{}, queueId:{}, offset:{}", cqExtUnit,
topic, queueId, request.getCommitLogOffset());
}
}
boolean result = this.putMessagePositionInfo(request.getCommitLogOffset(),
request.getMsgSize(), tagsCode, request.getConsumeQueueOffset());
if (result) {
if (this.defaultMessageStore.getMessageStoreConfig().getBrokerRole() == BrokerRole.SLAVE ||
this.defaultMessageStore.getMessageStoreConfig().isEnableDLegerCommitLog()) {
this.defaultMessageStore.getStoreCheckpoint().setPhysicMsgTimestamp(request.getStoreTimestamp());
}
this.defaultMessageStore.getStoreCheckpoint().setLogicsMsgTimestamp(request.getStoreTimestamp());
return;
} else {
}
}
this.defaultMessageStore.getRunningFlags().makeLogicsQueueError();
}
private boolean putMessagePositionInfo(final long offset, final int size, final long tagsCode,
final long cqOffset) {
if (offset + size <= this.maxPhysicOffset) {
log.warn("Maybe try to build consume queue repeatedly maxPhysicOffset={} phyOffset={}", maxPhysicOffset, offset);
return true;
}
this.byteBufferIndex.flip();
this.byteBufferIndex.limit(CQ_STORE_UNIT_SIZE);
this.byteBufferIndex.putLong(offset);
this.byteBufferIndex.putInt(size);
this.byteBufferIndex.putLong(tagsCode);
final long expectLogicOffset = cqOffset * CQ_STORE_UNIT_SIZE;
MappedFile mappedFile = this.mappedFileQueue.getLastMappedFile(expectLogicOffset);
if (mappedFile != null) {
if (mappedFile.isFirstCreateInQueue() && cqOffset != 0 && mappedFile.getWrotePosition() == 0) {
this.minLogicOffset = expectLogicOffset;
this.mappedFileQueue.setFlushedWhere(expectLogicOffset);
this.mappedFileQueue.setCommittedWhere(expectLogicOffset);
this.fillPreBlank(mappedFile, expectLogicOffset);
log.info("fill pre blank space " + mappedFile.getFileName() + " " + expectLogicOffset + " "
+ mappedFile.getWrotePosition());
}
if (cqOffset != 0) {
long currentLogicOffset = mappedFile.getWrotePosition() + mappedFile.getFileFromOffset();
if (expectLogicOffset < currentLogicOffset) {
return true;
}
if (expectLogicOffset != currentLogicOffset) {
}
}
this.maxPhysicOffset = offset + size;
return mappedFile.appendMessage(this.byteBufferIndex.array());
}
return false;
}
}
- ConsumeQueue保存的数据每一个条目共20个字节，分别为8字节的commitlog物理偏移量、4字节的消息长度、8字节tag hashcode。
CommitLogDispatcherBuildIndex
public class DefaultMessageStore implements MessageStore {
class CommitLogDispatcherBuildConsumeQueue implements CommitLogDispatcher {
@Override
public void dispatch(DispatchRequest request) {
final int tranType = MessageSysFlag.getTransactionValue(request.getSysFlag());
switch (tranType) {
case MessageSysFlag.TRANSACTION_NOT_TYPE:
case MessageSysFlag.TRANSACTION_COMMIT_TYPE:
DefaultMessageStore.this.putMessagePositionInfo(request);
break;
case MessageSysFlag.TRANSACTION_PREPARED_TYPE:
case MessageSysFlag.TRANSACTION_ROLLBACK_TYPE:
break;
}
}
}
class CommitLogDispatcherBuildIndex implements CommitLogDispatcher {
@Override
public void dispatch(DispatchRequest request) {
if (DefaultMessageStore.this.messageStoreConfig.isMessageIndexEnable()) {
DefaultMessageStore.this.indexService.buildIndex(request);
}
}
}
}
- dispatch执行DefaultMessageStore.this.indexService.buildIndex。
public class IndexService {
public void buildIndex(DispatchRequest req) {
IndexFile indexFile = retryGetAndCreateIndexFile();
if (indexFile != null) {
long endPhyOffset = indexFile.getEndPhyOffset();
DispatchRequest msg = req;
String topic = msg.getTopic();
String keys = msg.getKeys();
if (msg.getCommitLogOffset() < endPhyOffset) {
return;
}
final int tranType = MessageSysFlag.getTransactionValue(msg.getSysFlag());
switch (tranType) {
case MessageSysFlag.TRANSACTION_NOT_TYPE:
case MessageSysFlag.TRANSACTION_PREPARED_TYPE:
case MessageSysFlag.TRANSACTION_COMMIT_TYPE:
break;
case MessageSysFlag.TRANSACTION_ROLLBACK_TYPE:
return;
}
if (req.getUniqKey() != null) {
indexFile = putKey(indexFile, msg, buildKey(topic, req.getUniqKey()));
if (indexFile == null) {
log.error("putKey error commitlog {} uniqkey {}", req.getCommitLogOffset(), req.getUniqKey());
return;
}
}
if (keys != null && keys.length() > 0) {
String[] keyset = keys.split(MessageConst.KEY_SEPARATOR);
for (int i = 0; i < keyset.length; i++) {
String key = keyset[i];
if (key.length() > 0) {
indexFile = putKey(indexFile, msg, buildKey(topic, key));
if (indexFile == null) {
log.error("putKey error commitlog {} uniqkey {}", req.getCommitLogOffset(), req.getUniqKey());
return;
}
}
}
}
} else {
log.error("build index error, stop building index");
}
}
}
- IndexFile：IndexFile（索引文件）提供了一种可以通过key或时间区间来查询消息的方法。Index的文件名fileName是以创建时的时间戳命名的，固定的单个IndexFile文件大小约为400M，一个IndexFile可以保存 2000W个索引，IndexFile的底层存储设计为在文件系统中实现HashMap结构，故rocketmq的索引文件其底层实现为hash索引。
- 如果消息的properties中设置了UNIQ_KEY这个属性，就用 topic + “#” + UNIQ_KEY的value作为 key 来做写入操作。如果消息设置了KEYS属性（多个KEY以空格分隔），也会用 topic + “#” + KEY 来做索引。