文章目录
消息存储1.实时更新消息消费队列和索引文件1.1.转发到ConsumerQueue1.2.转发到Index2.消息队列和索引文件恢复2.1.存储文件加载2.1.1.加载commitLog文件2.1.2.加载消息消费队列2.1.3.加载索引文件2.2.正常恢复2.3.异常恢复消息存储
1.实时更新消息消费队列和索引文件
消息消费队文件、消息属性索引文件都是基于CommitLog文件构建的，当消息生产者提交的消息存储在CommitLog文件中，ConsumerQueue、IndexFile需要及时更新，否则消息无法及时被消费，根据消息属性查找消息也会出现较大延迟。RocketMQ通过开启一个线程ReputMessageService来准实时转发CommitLog文件更新事件，相应的任务处理器根据转发的消息及时更新ConsumerQueue、IndexFile文件。
构建消息消费队列和索引文件的时序图：
DefaultMessageStore#start
log.info("[SetReputOffset] maxPhysicalPosInLogicQueue={} clMinOffset={} clMaxOffset={} clConfirmedOffset={}",maxPhysicalPosInLogicQueue, mitLog.getMinOffset(), mitLog.getMaxOffset(), mitLog.getConfirmOffset());//设置CommitLog内存中最大偏移量this.reputMessageService.setReputFromOffset(maxPhysicalPosInLogicQueue);//启动消息分发服务线程this.reputMessageService.start();
ReputMessageService#run
DefaultMessageStore.log.info(this.getServiceName() + " service started");//线程状态为启动状态->每隔1毫秒就继续尝试推送消息到消息消费队列和索引文件while (!this.isStopped()) {try {Thread.sleep(1);//进行消息分发this.doReput();} catch (Exception e) {DefaultMessageStore.log.warn(this.getServiceName() + " service has exception. ", e);}}DefaultMessageStore.log.info(this.getServiceName() + " service end");
ReputMessageService#doReput
this.reputFromOffset = result.getStartOffset();for (int readSize = 0; readSize < result.getSize() && doNext; ) {//从result中循环遍历消息,一次读一条,创建DispatchRequest对象。DispatchRequest dispatchRequest =mitLog.checkMessageAndReturnSize(result.getByteBuffer(), false, false);int size = dispatchRequest.getBufferSize() == -1 ? dispatchRequest.getMsgSize() : dispatchRequest.getBufferSize();if (dispatchRequest.isSuccess()) {if (size > 0) {//---------------------------↓↓↓-------------------------//分发请求DefaultMessageStore.this.doDispatch(dispatchRequest);//当新消息达到 进行通知监听器进行处理if (BrokerRole.SLAVE != DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole()&& DefaultMessageStore.this.brokerConfig.isLongPollingEnable()&& DefaultMessageStore.this.messageArrivingListener != null) {//messageArrivingListener对新发送的消息进行监听DefaultMessageStore.this.messageArrivingListener.arriving(dispatchRequest.getTopic(),dispatchRequest.getQueueId(), dispatchRequest.getConsumeQueueOffset() + 1,dispatchRequest.getTagsCode(), dispatchRequest.getStoreTimestamp(),dispatchRequest.getBitMap(), dispatchRequest.getPropertiesMap());}//更新消息分发偏移量 this.reputFromOffset += size;readSize += size;if (DefaultMessageStore.this.getMessageStoreConfig().getBrokerRole() == BrokerRole.SLAVE) {DefaultMessageStore.this.storeStatsService.getSinglePutMessageTopicTimesTotal(dispatchRequest.getTopic()).add(1);DefaultMessageStore.this.storeStatsService.getSinglePutMessageTopicSizeTotal(dispatchRequest.getTopic()).add(dispatchRequest.getMsgSize());}} ....}
DispatchRequest
String topic; //消息主题名称int queueId; //消息队列IDlong commitLogOffset;//消息物理偏移量int msgSize;//消息长度long tagsCode;//消息过滤tag hashCodelong storeTimestamp;//消息存储时间戳long consumeQueueOffset;//消息队列偏移量String keys;//消息索引keyboolean success;//是否成功解析到完整的消息String uniqKey;//消息唯一键int sysFlag;//消息系统标记long preparedTransactionOffset;//消息预处理事务偏移量Map<String, String> propertiesMap;//消息属性byte[] bitMap;//位图
DefaultMessageStore#doDispatch
for (CommitLogDispatcher dispatcher : this.dispatcherList) {dispatcher.dispatch(req);}
1.1.转发到ConsumerQueue
CommitLogDispatcherBuildConsumeQueue#dispatch
final int tranType = MessageSysFlag.getTransactionValue(request.getSysFlag());switch (tranType) {case MessageSysFlag.TRANSACTION_NOT_TYPE:case MessageSysFlag.TRANSACTION_COMMIT_TYPE://消息分发DefaultMessageStore.this.putMessagePositionInfo(request);break;case MessageSysFlag.TRANSACTION_PREPARED_TYPE:case MessageSysFlag.TRANSACTION_ROLLBACK_TYPE:break;}
DefaultMessageStore#putMessagePositionInfo
//根据请求中的主题和队列ID获得消费队列ConsumeQueue cq = this.findConsumeQueue(dispatchRequest.getTopic(), dispatchRequest.getQueueId());//消息队列分发消息cq.putMessagePositionInfoWrapper(dispatchRequest);
ConsumeQueue#putMessagePositionInfoWrapper
boolean result = this.putMessagePositionInfo(request.getCommitLogOffset(),
ConsumeQueue#putMessagePositionInfo
private boolean putMessagePositionInfo(final long offset, //commitLog偏移量final int size, //消息体大小final long tagsCode,//消息tags的hashCodefinal long cqOffset) {//写入consumerqueue的偏移量//依次将消息偏移量、消息长度、tag写入到ByteBuffer中this.byteBufferIndex.flip();this.byteBufferIndex.limit(CQ_STORE_UNIT_SIZE);this.byteBufferIndex.putLong(offset);this.byteBufferIndex.putInt(size);this.byteBufferIndex.putLong(tagsCode);//计算期望插入ConsumerQueue的文件位置final long expectLogicOffset = cqOffset * CQ_STORE_UNIT_SIZE;//获得内存映射文件MappedFile mappedFile = this.mappedFileQueue.getLastMappedFile(expectLogicOffset);if(mappedFile != null){//如果文件是新建的->进行填充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());}//将消息追加到内存映射文件,异步输盘->整个过程都是基于MappedFilereturn mappedFile.appendMessage(this.byteBufferIndex.array());}}
1.2.转发到Index
核心实现类是IndexService，存储Index文件的封装类是IndexFile。
IndexFile
// 每个 hash 槽所占的字节数private static int hashSlotSize = 4;//每个indexFile条目所占用字节数private static int indexSize = 20;//用来验证是否是一个有效索引private static int invalidIndex = 0;//indexFile中hash槽的个数private final int hashSlotNum;//indexFile中包含的条目数private final int indexNum;//对应的MappedFileprivate final MappedFile mappedFile;//文件传输通道private final FileChannel fileChannel;//pageCacheprivate final MappedByteBuffer mappedByteBuffer;//每一个IndexFile的头部信息private final IndexHeader indexHeader;
消息分发到索引文件的时序图：
CommitLogDispatcherBuildIndex#dispatch
//开启文件索引if (DefaultMessageStore.this.messageStoreConfig.isMessageIndexEnable()) {//构建索引DefaultMessageStore.this.indexService.buildIndex(request);}
IndexService#buildIndex
//创建或获取索引文件IndexFile indexFile = retryGetAndCreateIndexFile();if (indexFile != null) {//获得文件最大物理偏移量long endPhyOffset = indexFile.getEndPhyOffset();DispatchRequest msg = req;//消息主题String topic = msg.getTopic();//消息keyString keys = msg.getKeys();//如果该消息的物理偏移量小于索引文件中的最大物理偏移量,则说明是重复数据,忽略本次索引构建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;}//如果消息ID不为空,则添加到Hash索引中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;}}//构建索引key,RocketMQ支持为同一个消息建立多个索引,多个索引键空格隔开.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");}
IndexService#putKey
for (boolean ok = indexFile.putKey(idxKey, msg.getCommitLogOffset(), msg.getStoreTimestamp()); !ok; ) {log.warn("Index file [" + indexFile.getFileName() + "] is full, trying to create another one");//重试-获取或者创建Index文件indexFile = retryGetAndCreateIndexFile();if (null == indexFile) {return null;}ok = indexFile.putKey(idxKey, msg.getCommitLogOffset(), msg.getStoreTimestamp());}return indexFile;
IndexFile#putKey
public boolean putKey(final String key, final long phyOffset, //消息存储在commitLog的偏移量final long storeTimestamp) {//消息存入commitLog的时间戳//如果IndexFile存储的条目数小于最大条目数限制-允许存储//否则表示存储失败-返回falseif (this.indexHeader.getIndexCount() < this.indexNum) {int keyHash = indexKeyHashMethod(key);int slotPos = keyHash % this.hashSlotNum;//根据keyHash和hash槽数量取模->得到该key在槽中的下标int absSlotPos = IndexHeader.INDEX_HEADER_SIZE + slotPos * hashSlotSize;FileLock fileLock = null;try {// fileLock = this.fileChannel.lock(absSlotPos, hashSlotSize,// false);int slotValue = this.mappedByteBuffer.getInt(absSlotPos);if (slotValue <= invalidIndex || slotValue > this.indexHeader.getIndexCount()) {slotValue = invalidIndex;}long timeDiff = storeTimestamp - this.indexHeader.getBeginTimestamp();timeDiff = timeDiff / 1000;if (this.indexHeader.getBeginTimestamp() <= 0) {timeDiff = 0;} else if (timeDiff > Integer.MAX_VALUE) {timeDiff = Integer.MAX_VALUE;} else if (timeDiff < 0) {timeDiff = 0;}int absIndexPos =IndexHeader.INDEX_HEADER_SIZE + this.hashSlotNum * hashSlotSize+ this.indexHeader.getIndexCount() * indexSize;this.mappedByteBuffer.putInt(absIndexPos, keyHash);this.mappedByteBuffer.putLong(absIndexPos + 4, phyOffset);this.mappedByteBuffer.putInt(absIndexPos + 4 + 8, (int) timeDiff);this.mappedByteBuffer.putInt(absIndexPos + 4 + 8 + 4, slotValue);this.mappedByteBuffer.putInt(absSlotPos, this.indexHeader.getIndexCount());if (this.indexHeader.getIndexCount() <= 1) {this.indexHeader.setBeginPhyOffset(phyOffset);this.indexHeader.setBeginTimestamp(storeTimestamp);}if (invalidIndex == slotValue) {this.indexHeader.incHashSlotCount();}this.indexHeader.incIndexCount();this.indexHeader.setEndPhyOffset(phyOffset);this.indexHeader.setEndTimestamp(storeTimestamp);return true;} catch (Exception e) {log.error("putKey exception, Key: " + key + " KeyHashCode: " + key.hashCode(), e);} finally {if (fileLock != null) {try {fileLock.release();} catch (IOException e) {log.error("Failed to release the lock", e);}}}} else {log.warn("Over index file capacity: index count = " + this.indexHeader.getIndexCount()+ "; index max num = " + this.indexNum);}return false;}
这里涉及到索引文件的存储格式和查找，暂不在此做解析。
2.消息队列和索引文件恢复
由于RocketMQ存储首先将消息全量存储在CommitLog文件中，然后异步生成转发任务更新ConsumerQueue和Index文件。如果消息成功存储到CommitLog文件中，转发任务未成功执行，此时消息服务器Broker由于某种原因宕机，导致CommitLog、ConsumerQueue、IndexFile文件数据不一致。如果不加以人工修复的话，会有一部分消息即便在CommitLog中文件中存在，但由于没有转发到ConsumerQueue，这部分消息将永远不被消费者消费。
2.1.存储文件加载
判断上一次是否是异常退出——实现机制是Broker在启动时会创建abort文件，在退出时会通过JVM钩子函数删除abort文件。如果下次启动时发现abort仍然存在，则说明Broker上次退出是异常的，CommitLog和ConsumerQueue数据可能存在不一致的情况，需要进行文件修复。
//检查上次退出是否异常boolean lastExitOK = !this.isTempFileExist();↓↓↓private boolean isTempFileExist() {//尝试获取abort临时文件String fileName = StorePathConfigHelper.getAbortFile(this.messageStoreConfig.getStorePathRootDir());File file = new File(fileName);//如果存在即说明上次退出异常return file.exists();}
DefaultMessageStore#load
//加载结果boolean result = true;try {//判断是否是正常退出boolean lastExitOK = !this.isTempFileExist();log.info("last shutdown {}", lastExitOK ? "normally" : "abnormally");//1.加载commitLog文件result = result && mitLog.load();//2.加载consumeQueue文件result = result && this.loadConsumeQueue();if (result) {//加载store目录下的所有存储文件this.storeCheckpoint =new StoreCheckpoint(StorePathConfigHelper.getStoreCheckpoint(this.messageStoreConfig.getStorePathRootDir()));//3.加载index文件this.indexService.load(lastExitOK);//4.根据broker是否异常退出采取不同的恢复策略this.recover(lastExitOK);log.info("load over, and the max phy offset = {}", this.getMaxPhyOffset());if (null != scheduleMessageService) {result = this.scheduleMessageService.load();}}} catch (Exception e) {log.error("load exception", e);result = false;}//如果加载失败->关闭MappedFile分配服务线程if (!result) {this.allocateMappedFileService.shutdown();}return result;
2.1.1.加载commitLog文件
commitLog.load()→mappedFileQueue.load()
MappedFileQueue#load
//根据文件目录获取文件File dir = new File(this.storePath);//获得文件数组File[] ls = dir.listFiles();if (ls != null) {return doLoad(Arrays.asList(ls));}return true;
MappedFileQueue#doLoad
//文件进行排序files.sort(paring(File::getName));//遍历文件列表for (File file : files) {//如果文件大小与配置文件不一致 退出if (file.length() != this.mappedFileSize) {log.warn(file + "\t" + file.length()+ " length not matched message store config value, ignore it");return true;}try {//创建映射文件MappedFile mappedFile = new MappedFile(file.getPath(), mappedFileSize);//设置映射文件的指针mappedFile.setWrotePosition(this.mappedFileSize);mappedFile.setFlushedPosition(this.mappedFileSize);mappedFile.setCommittedPosition(this.mappedFileSize);//将映射文件加到队列中this.mappedFiles.add(mappedFile);log.info("load " + file.getPath() + " OK");} catch (IOException e) {log.error("load file " + file + " error", e);return false;}}return true;
2.1.2.加载消息消费队列
DefaultMessageStore#loadConsumeQueue
//获得消费队列目录File dirLogic = new File(StorePathConfigHelper.getStorePathConsumeQueue(this.messageStoreConfig.getStorePathRootDir()));File[] fileTopicList = dirLogic.listFiles();if (fileTopicList != null) {//遍历消费队列目录for (File fileTopic : fileTopicList) {//获得子目录名称，即为topic名称String topic = fileTopic.getName();//遍历子目录下的消费队列文件File[] fileQueueIdList = fileTopic.listFiles();if (fileQueueIdList != null) {//遍历文件for (File fileQueueId : fileQueueIdList) {int queueId;try {queueId = Integer.parseInt(fileQueueId.getName());} catch (NumberFormatException e) {continue;}//创建消费目录并加载到内存中ConsumeQueue logic = new ConsumeQueue(topic,queueId,StorePathConfigHelper.getStorePathConsumeQueue(this.messageStoreConfig.getStorePathRootDir()),this.getMessageStoreConfig().getMappedFileSizeConsumeQueue(),this);//存入consumeQueueTable消息队列表this.putConsumeQueue(topic, queueId, logic);if (!logic.load()) {return false;}}}}}//打印加载消费队列成功日志log.info("load logics queue all over, OK");return true;
2.1.3.加载索引文件
IndexService#load
//根据索引文件目录加载文件File dir = new File(this.storePath);//获取文件数组File[] files = dir.listFiles();if (files != null) {// ascending order//文件排序Arrays.sort(files);//文件遍历for (File file : files) {try {//创建索引文件IndexFile f = new IndexFile(file.getPath(), this.hashSlotNum, this.indexNum, 0, 0);//加载f.load();//如果异常退出if (!lastExitOK) {//索引文件上次刷盘时间大于检测点时间戳if (f.getEndTimestamp() > this.defaultMessageStore.getStoreCheckpoint().getIndexMsgTimestamp()) {//该文件立刻删除f.destroy(0);continue;}}log.info("load index file OK, " + f.getFileName());//将符合条件索引文件加入列表this.indexFileList.add(f);} catch (IOException e) {log.error("load file {} error", file, e);return false;} catch (NumberFormatException e) {log.error("load file {} error", file, e);}}}return true;
所有文件均加载完毕，进行文件恢复，根据Broker是否正常退出执行不同的恢复策略。
DefaultMessageStore#recover
long maxPhyOffsetOfConsumeQueue = this.recoverConsumeQueue();if (lastExitOK) {//正常恢复mitLog.recoverNormally(maxPhyOffsetOfConsumeQueue);} else {//异常恢复mitLog.recoverAbnormally(maxPhyOffsetOfConsumeQueue);}//在CommitLog中保存每个消息消费队列当前的存储逻辑偏移量this.recoverTopicQueueTable();
DefaultMessageStore#recoverConsumeQueue
long maxPhysicOffset = -1;for (ConcurrentMap<Integer, ConsumeQueue> maps : this.consumeQueueTable.values()) {for (ConsumeQueue logic : maps.values()) {//覆盖消息队列logic.recover();if (logic.getMaxPhysicOffset() > maxPhysicOffset) {maxPhysicOffset = logic.getMaxPhysicOffset();}}}return maxPhysicOffset;
ConsumeQueue#recover
//获取消息队列的所有映射文件final List<MappedFile> mappedFiles = this.mappedFileQueue.getMappedFiles();if (!mappedFiles.isEmpty()) {int index = mappedFiles.size() - 3;//从倒数第三个文件开始if (index < 0) {index = 0;}//consumerQueue逻辑大小int mappedFileSizeLogics = this.mappedFileSize;//consumerQueue对应的映射文件MappedFile mappedFile = mappedFiles.get(index);//映射文件对应的ByteBufferByteBuffer byteBuffer = mappedFile.sliceByteBuffer();//处理的offset-默认从consumerQueue中存放的第一个条目开始long processOffset = mappedFile.getFileFromOffset();long mappedFileOffset = 0;long maxExtAddr = 1;while (true) {//循环验证consumerQueue包含条目的有效性for (int i = 0; i < mappedFileSizeLogics; i += CQ_STORE_UNIT_SIZE) {//读取条目内容//commitLog物理偏移量long offset = byteBuffer.getLong();//消息总长度int size = byteBuffer.getInt();//tag哈希值long tagsCode = byteBuffer.getLong();//offset大于0并且size大于0说明该条目是一个有效的if (offset >= 0 && size > 0) {mappedFileOffset = i + CQ_STORE_UNIT_SIZE;this.maxPhysicOffset = offset + size;if (isExtAddr(tagsCode)) {maxExtAddr = tagsCode;}} else {log.info("recover current consume queue file over, " + mappedFile.getFileName() + " "+ offset + " " + size + " " + tagsCode);break;}}//如果该consumeQueue中包含的条目全部有效则继续验证下一个文件-index++if (mappedFileOffset == mappedFileSizeLogics) {index++;//验证完毕-退出-开始恢复消息队列文件if (index >= mappedFiles.size()) {log.info("recover last consume queue file over, last mapped file "+ mappedFile.getFileName());break;} else {//获取下一个映射文件-继续循环检查mappedFile = mappedFiles.get(index);byteBuffer = mappedFile.sliceByteBuffer();processOffset = mappedFile.getFileFromOffset();mappedFileOffset = 0;log.info("recover next consume queue file, " + mappedFile.getFileName());}} else {log.info("recover current consume queue queue over " + mappedFile.getFileName() + " "+ (processOffset + mappedFileOffset));break;}}//当前consumerQueue有效偏移量processOffset += mappedFileOffset;//设置flushWhere、committedWhere为有效偏移量this.mappedFileQueue.setFlushedWhere(processOffset);this.mappedFileQueue.setCommittedWhere(processOffset);//删除冗余无效的consumerQueue文件this.mappedFileQueue.truncateDirtyFiles(processOffset);if (isExtReadEnable()) {this.consumeQueueExt.recover();log.info("Truncate consume queue extend file by max {}", maxExtAddr);this.consumeQueueExt.truncateByMaxAddress(maxExtAddr);}}
DefaultMessageStore#recoverTopicQueueTable
HashMap<String/* topic-queueid */, Long/* offset */> table = new HashMap<String, Long>(1024);//CommitLog最小偏移量long minPhyOffset = mitLog.getMinOffset();//遍历消费队列,将消费队列保存在CommitLog中for (ConcurrentMap<Integer, ConsumeQueue> maps : this.consumeQueueTable.values()) {for (ConsumeQueue logic : maps.values()) {String key = logic.getTopic() + "-" + logic.getQueueId();table.put(key, logic.getMaxOffsetInQueue());logic.correctMinOffset(minPhyOffset);}}//覆盖主题消息队列信息mitLog.setTopicQueueTable(table);
2.2.正常恢复
CommitLog恢复过程与ConsumeQueue恢复过程极其相似。
CommitLog#recoverNormally
boolean checkCRCOnRecover = this.defaultMessageStore.getMessageStoreConfig().isCheckCRCOnRecover();//获取映射文件列表final List<MappedFile> mappedFiles = this.mappedFileQueue.getMappedFiles();if (!mappedFiles.isEmpty()) {//Broker正常停止再重启时,从倒数第三个开始恢复,如果不足3个文件,则从第一个文件开始恢复。int index = mappedFiles.size() - 3;if (index < 0) {index = 0;}MappedFile mappedFile = mappedFiles.get(index);ByteBuffer byteBuffer = mappedFile.sliceByteBuffer();long processOffset = mappedFile.getFileFromOffset();//代表当前已校验通过的offsetlong mappedFileOffset = 0;while (true) {//检查消息并返回分发请求DispatchRequest dispatchRequest = this.checkMessageAndReturnSize(byteBuffer, checkCRCOnRecover);//消息长度int size = dispatchRequest.getMsgSize();//查找结果为true,并且消息长度大于0,表示消息正确.mappedFileOffset向前移动本消息长度if (dispatchRequest.isSuccess() && size > 0) {mappedFileOffset += size;}//如果查找结果为true且消息长度等于0,表示已到该文件末尾,如果还有下一个文件,则重置processOffset和MappedFileOffset重复查找下一个文件,否则跳出循环。else if (dispatchRequest.isSuccess() && size == 0) {index++;//完成检查-退出循环if (index >= mappedFiles.size()) {// Current branch can not happenlog.info("recover last 3 physics file over, last mapped file " + mappedFile.getFileName());break;} else {//获取并检查下一个映射文件mappedFile = mappedFiles.get(index);byteBuffer = mappedFile.sliceByteBuffer();processOffset = mappedFile.getFileFromOffset();//重置检查offsetmappedFileOffset = 0;log.info("recover next physics file, " + mappedFile.getFileName());}}// Intermediate file read errorelse if (!dispatchRequest.isSuccess()) {log.info("recover physics file end, " + mappedFile.getFileName());break;}}//更新MappedFileQueue的flushedWhere和committedWhere指针processOffset += mappedFileOffset;this.mappedFileQueue.setFlushedWhere(processOffset);this.mappedFileQueue.setCommittedWhere(processOffset);//删除之后的冗余无效的文件this.mappedFileQueue.truncateDirtyFiles(processOffset);// Clear ConsumeQueue redundant dataif (maxPhyOffsetOfConsumeQueue >= processOffset) {log.warn("maxPhyOffsetOfConsumeQueue({}) >= processOffset({}), truncate dirty logic files", maxPhyOffsetOfConsumeQueue, processOffset);this.defaultMessageStore.truncateDirtyLogicFiles(processOffset);}} else {// Commitlog case files are deletedlog.warn("The commitlog files are deleted, and delete the consume queue files");this.mappedFileQueue.setFlushedWhere(0);this.mappedFileQueue.setCommittedWhere(0);this.defaultMessageStore.destroyLogics();
MappedFileQueue#truncateDirtyFiles
//即将移除文件List<MappedFile> willRemoveFiles = new ArrayList<MappedFile>();//遍历映射文件for (MappedFile file : this.mappedFiles) {long fileTailOffset = file.getFileFromOffset() + this.mappedFileSize;//文件尾部的偏移量大于offsetif (fileTailOffset > offset) {//offset大于文件的起始偏移量if (offset >= file.getFileFromOffset()) {//更新wrotePosition、committedPosition、flushedPosistionfile.setWrotePosition((int) (offset % this.mappedFileSize));file.setCommittedPosition((int) (offset % this.mappedFileSize));file.setFlushedPosition((int) (offset % this.mappedFileSize));} else {//offset小于文件的起始偏移量,说明该文件是有效文件后面创建的,释放mappedFile占用内存,删除文件file.destroy(1000);willRemoveFiles.add(file);}}}this.deleteExpiredFile(willRemoveFiles);
2.3.异常恢复
Broker异常停止文件恢复的实现为CommitLog#recoverAbnormally。异常文件恢复步骤与正常停止文件恢复流程基本相同，其主要差别有两个。首先，正常停止默认从倒数第三个文件开始进行恢复，而异常停止则需要从最后一个文件往前走，找到第一个消息存储正常的文件。其次，如果CommitLog目录没有消息文件，如果消息消费队列目录下存在文件，则需要销毁。
CommitLog#recoverAbnormally
boolean checkCRCOnRecover = this.defaultMessageStore.getMessageStoreConfig().isCheckCRCOnRecover();final List<MappedFile> mappedFiles = this.mappedFileQueue.getMappedFiles();if (!mappedFiles.isEmpty()) {// Looking beginning to recover from which file//从最后一个文件开始检查int index = mappedFiles.size() - 1;MappedFile mappedFile = null;for (; index >= 0; index--) {mappedFile = mappedFiles.get(index);//如果该文件是一个正确的文件if (this.isMappedFileMatchedRecover(mappedFile)) {log.info("recover from this mapped file " + mappedFile.getFileName());break;}}//恰好正确文件是第一个-<index-->之后index小于0if (index < 0) {//需要将索引校正为0index = 0;//取出第一个文件-从第一个文件开始恢复mappedFile = mappedFiles.get(index);}ByteBuffer byteBuffer = mappedFile.sliceByteBuffer();long processOffset = mappedFile.getFileFromOffset();long mappedFileOffset = 0;while (true) {//验证消息是否合法DispatchRequest dispatchRequest = this.checkMessageAndReturnSize(byteBuffer, checkCRCOnRecover);int size = dispatchRequest.getMsgSize();//消息校验成功if (dispatchRequest.isSuccess()) {// 正qif (size > 0) {mappedFileOffset += size;//将消息转发到消息消费队列和索引文件if (this.defaultMessageStore.getMessageStoreConfig().isDuplicationEnable()) {if (dispatchRequest.getCommitLogOffset() < this.defaultMessageStore.getConfirmOffset()) {this.defaultMessageStore.doDispatch(dispatchRequest);}} else {this.defaultMessageStore.doDispatch(dispatchRequest);}}//到达文件尾else if (size == 0) {index++;//文件索引向后移动//文件遍历完毕if (index >= mappedFiles.size()) {// The current branch under normal circumstances should// not happenlog.info("recover physics file over, last mapped file " + mappedFile.getFileName());break;} else {//获取下一个的映射文件mappedFile = mappedFiles.get(index);byteBuffer = mappedFile.sliceByteBuffer();processOffset = mappedFile.getFileFromOffset();mappedFileOffset = 0;log.info("recover next physics file, " + mappedFile.getFileName());}}} else {log.info("recover physics file end, " + mappedFile.getFileName() + " pos=" + byteBuffer.position());break;}}//当前有效偏移量processOffset += mappedFileOffset;//设置flushWhere、committedWhere为有效偏移量this.mappedFileQueue.setFlushedWhere(processOffset);this.mappedFileQueue.setCommittedWhere(processOffset);//清除冗余的数据this.mappedFileQueue.truncateDirtyFiles(processOffset);//清除冗余的数据if (maxPhyOffsetOfConsumeQueue >= processOffset) {log.warn("maxPhyOffsetOfConsumeQueue({}) >= processOffset({}), truncate dirty logic files", maxPhyOffsetOfConsumeQueue, processOffset);this.defaultMessageStore.truncateDirtyLogicFiles(processOffset);}}//文件被删除else {log.warn("The commitlog files are deleted, and delete the consume queue files");//flushWhere committedWhere重置this.mappedFileQueue.setFlushedWhere(0);this.mappedFileQueue.setCommittedWhere(0);//销毁消息队列文件this.defaultMessageStore.destroyLogics();}
本文仅作为个人学习使用，如有不足或错误请指正！
