flume高并发优化——（8）多文件source扩展断点续传

在很多情况下，我们为了不丢失数据，一般都会为数据收集端扩展断点续传，而随着公司日志系统的完善，我们在原有的基础上开发了断点续传的功能，以下是思路，大家共同讨论：

核心流程图：

源码：

/*
* 作者：许恕
* 时间：2016年5月3日
* 功能：实现tail 某目录下的所有符合正则条件的文件
* Email：xvshu1@163.com
* To detect all files in a folder
*/

package org.apache.flume.source;

import com.google.common.base.Preconditions;
import com.google.common.util.concurrent.ThreadFactoryBuilder;
import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.EventDrivenSource;
import org.apache.flume.SystemClock;
import org.apache.flume.channel.ChannelProcessor;
import org.apache.flume.conf.Configurable;
import org.apache.flume.event.EventBuilder;
import org.apache.flume.instrumentation.SourceCounter;
import org.apache.flume.tools.HostUtils;
import org.mortbay.util.ajax.JSON;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.*;
import java.nio.charset.Charset;
import java.util.*;
import java.util.concurrent.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

/**
*  step：
*    1,config one path
*    2,find all file with RegExp
*    3,tail one children file
*    4,batch to channal
*
*  demo:
*    demo.sources.s1.type = org.apache.flume.source.ExecTailSource
*    demo.sources.s1.filepath=/export/home/tomcat/logs/auth.el.net/
*    demo.sources.s1.filenameRegExp=(.log{1})$
*    demo.sources.s1.tailing=true
*    demo.sources.s1.readinterval=300
*    demo.sources.s1.startAtBeginning=false
*    demo.sources.s1.restart=true
*/
public class ExecTailSource extends AbstractSource implements EventDrivenSource,
Configurable {

private static final Logger logger = LoggerFactory
.getLogger(ExecTailSource.class);

private SourceCounter sourceCounter;
private ExecutorService executor;
private List<ExecRunnable> listRuners;
private List<Future<?>> listFuture;
private long restartThrottle;
private boolean restart = true;
private boolean logStderr;
private Integer bufferCount;
private long batchTimeout;
private Charset charset;
private String filepath;
private String filenameRegExp;
private boolean tailing;
private Integer readinterval;
private boolean startAtBeginning;
private boolean contextIsJson;
private String fileWriteJson;
private Long flushTime;
private boolean contextIsFlumeLog;

@Override
public void start() {
logger.info("=start=> flume tail source start begin time:"+new Date().toString());
logger.info("ExecTail source starting with filepath:{}", filepath);

List<String> listFiles = getFileList(filepath);
if(listFiles==null || listFiles.isEmpty()){
Preconditions.checkState(listFiles != null && !listFiles.isEmpty(),
"The filepath's file not have fiels with filenameRegExp");
}

Properties prop=null;

try{
prop = new Properties();//属性集合对象
FileInputStream fis = new FileInputStream(fileWriteJson);//属性文件流
prop.load(fis);
}catch(Exception ex){
logger.error("==>",ex);
}

executor = Executors.newFixedThreadPool(listFiles.size());

listRuners = new ArrayList<ExecRunnable>();
listFuture = new ArrayList<Future<?>>();

logger.info("files size is {} ", listFiles.size());
// FIXME: Use a callback-like executor / future to signal us upon failure.
for(String oneFilePath : listFiles){
ExecRunnable runner = new ExecRunnable(getChannelProcessor(), sourceCounter,
restart, restartThrottle, logStderr, bufferCount, batchTimeout,
charset,oneFilePath,tailing,readinterval,startAtBeginning,contextIsJson,
prop,fileWriteJson,flushTime,contextIsFlumeLog);
listRuners.add(runner);
Future<?> runnerFuture = executor.submit(runner);
listFuture.add(runnerFuture);
logger.info("{} is begin running",oneFilePath);
}

/*
* NB: This comes at the end rather than the beginning of the method because
* it sets our state to running. We want to make sure the executor is alive
* and well first.
*/
sourceCounter.start();
super.start();
logger.info("=start=> flume tail source start end time:"+new Date().toString());
logger.debug("ExecTail source started");
}

@Override
public void stop() {

logger.info("=stop=> flume tail source stop begin time:"+new Date().toString());
if(listRuners !=null && !listRuners.isEmpty()){
for(ExecRunnable oneRunner : listRuners){
if(oneRunner != null) {
oneRunner.setRestart(false);
oneRunner.kill();
}
}
}

if(listFuture !=null && !listFuture.isEmpty()){
for(Future<?> oneFuture : listFuture){
if (oneFuture != null) {
logger.debug("Stopping ExecTail runner");
oneFuture.cancel(true);
logger.debug("ExecTail runner stopped");
}
}
}

executor.shutdown();
while (!executor.isTerminated()) {
logger.debug("Waiting for ExecTail executor service to stop");
try {
executor.awaitTermination(500, TimeUnit.MILLISECONDS);
} catch (InterruptedException e) {
logger.debug("Interrupted while waiting for ExecTail executor service "
+ "to stop. Just exiting.");
Thread.currentThread().interrupt();
}
}

sourceCounter.stop();
super.stop();
logger.info("=stop=> flume tail source stop end time:"+new Date().toString());

}

@Override
public void configure(Context context) {

filepath = context.getString("filepath");
Preconditions.checkState(filepath != null,
"The parameter filepath must be specified");
logger.info("The parameter filepath is {}" ,filepath);

filenameRegExp = context.getString("filenameRegExp");
Preconditions.checkState(filenameRegExp != null,
"The parameter filenameRegExp must be specified");
logger.info("The parameter filenameRegExp is {}" ,filenameRegExp);

contextIsJson= context.getBoolean(ExecSourceConfigurationConstants.CONFIG_CONTEXTISJSON_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_CONTEXTISJSON);

contextIsFlumeLog=context.getBoolean(ExecSourceConfigurationConstants.CONFIG_CONTEXTISFLUMELOG_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_CONTEXTISFLUMELOG);

fileWriteJson= context.getString(ExecSourceConfigurationConstants.CONFIG_FILEWRITEJSON_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_FILEWRITEJSON);

flushTime= context.getLong(ExecSourceConfigurationConstants.CONFIG_FLUSHTIME_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_FLUSHTIME);

restartThrottle = context.getLong(ExecSourceConfigurationConstants.CONFIG_RESTART_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_RESTART_THROTTLE);

tailing = context.getBoolean(ExecSourceConfigurationConstants.CONFIG_TAILING_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_ISTAILING_TRUE);

readinterval=context.getInteger(ExecSourceConfigurationConstants.CONFIG_READINTERVAL_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_READINTERVAL);

startAtBeginning=context.getBoolean(ExecSourceConfigurationConstants.CONFIG_STARTATBEGINNING_THROTTLE,
ExecSourceConfigurationConstants.DEFAULT_STARTATBEGINNING);

restart = context.getBoolean(ExecSourceConfigurationConstants.CONFIG_RESTART,
ExecSourceConfigurationConstants.DEFAULT_RESTART_TRUE);

logStderr = context.getBoolean(ExecSourceConfigurationConstants.CONFIG_LOG_STDERR,
ExecSourceConfigurationConstants.DEFAULT_LOG_STDERR);

bufferCount = context.getInteger(ExecSourceConfigurationConstants.CONFIG_BATCH_SIZE,
ExecSourceConfigurationConstants.DEFAULT_BATCH_SIZE);

batchTimeout = context.getLong(ExecSourceConfigurationConstants.CONFIG_BATCH_TIME_OUT,
ExecSourceConfigurationConstants.DEFAULT_BATCH_TIME_OUT);

charset = Charset.forName(context.getString(ExecSourceConfigurationConstants.CHARSET,
ExecSourceConfigurationConstants.DEFAULT_CHARSET));

if (sourceCounter == null) {
sourceCounter = new SourceCounter(getName());
}
}

/**
* 获取指定路径下的所有文件列表
*
* @param dir 要查找的目录
* @return
*/
public  List<String> getFileList(String dir) {
List<String> listFile = new ArrayList<String>();
File dirFile = new File(dir);
//如果不是目录文件，则直接返回
if (dirFile.isDirectory()) {
//获得文件夹下的文件列表，然后根据文件类型分别处理
File[] files = dirFile.listFiles();
if (null != files && files.length > 0) {
//根据时间排序
Arrays.sort(files, new Comparator<File>() {
public int compare(File f1, File f2) {
return (int) (f1.lastModified() - f2.lastModified());
}

public boolean equals(Object obj) {
return true;
}
});
for (File file : files) {
//如果不是目录，直接添加
if (!file.isDirectory()) {
String oneFileName = file.getName();
if(match(filenameRegExp,oneFileName)){
listFile.add(file.getAbsolutePath());
logger.info("filename:{} is pass",oneFileName);
}
} else {
//对于目录文件，递归调用
listFile.addAll(getFileList(file.getAbsolutePath()));
}
}
}
}else{
logger.info("FilePath:{} is not Directory",dir);
}
return listFile;
}

/**
* @param regex
* 正则表达式字符串
* @param str
* 要匹配的字符串
* @return 如果str 符合 regex的正则表达式格式,返回true, 否则返回 false;
*/
private boolean match(String regex, String str) {
Pattern pattern = Pattern.compile(regex);
Matcher matcher = pattern.matcher(str);
return matcher.find();
}

private static class ExecRunnable implements Runnable {

public ExecRunnable( ChannelProcessor channelProcessor,
SourceCounter sourceCounter, boolean restart, long restartThrottle,
boolean logStderr, int bufferCount, long batchTimeout,
Charset charset, String filepath,
boolean tailing,Integer readinterval,
boolean startAtBeginning,boolean contextIsJson,
Properties prop,String fileWriteJson,Long flushTime,
boolean contextIsFlumeLog) {

this.channelProcessor = channelProcessor;
this.sourceCounter = sourceCounter;
this.restartThrottle = restartThrottle;
this.bufferCount = bufferCount;
this.batchTimeout = batchTimeout;
this.restart = restart;
this.logStderr = logStderr;
this.charset = charset;
this.filepath=filepath;
this.logfile=new File(filepath);
this.tailing=tailing;
this.readinterval=readinterval;
this.startAtBeginning=startAtBeginning;
this.contextIsJson=contextIsJson;
this.prop = prop;
this.fileWriteJson=fileWriteJson;
this.flushTime=flushTime;
this.contextIsFlumeLog=contextIsFlumeLog;
}

private final ChannelProcessor channelProcessor;
private final SourceCounter sourceCounter;
private volatile boolean restart;
private final long restartThrottle;
private final int bufferCount;
private long batchTimeout;
private final boolean logStderr;
private final Charset charset;
private SystemClock systemClock = new SystemClock();
private Long lastPushToChannel = systemClock.currentTimeMillis();
ScheduledExecutorService timedFlushService;
ScheduledFuture<?> future;
private String filepath;
private boolean contextIsJson;
private Properties prop;
private long timepoint;
private String fileWriteJson;
private Long flushTime;

/**
* 当读到文件结尾后暂停的时间间隔
*/
private long readinterval = 500;

/**
* 设置日志文件
*/
private File logfile;

/**
* 设置是否从头开始读
*/
private boolean startAtBeginning = false;

/**
* 设置tail运行标记
*/
private boolean tailing = false;

private boolean contextIsFlumeLog=false;

private static String getDomain(String filePath){
String[] strs = filePath.split("/");
String domain ;
domain=strs[strs.length-2];
if(domain==null || domain.isEmpty()){
domain=filePath;
}
return domain;
}

@Override
public void run() {
do {
logger.info("=run=> flume tail source run start time:"+new Date().toString());
timepoint=System.currentTimeMillis();
Long filePointer = 0l;
if (this.startAtBeginning) { //判断是否从头开始读文件
filePointer = 0l;
} else {
if(prop!=null || prop.contains(filepath)){

try {
filePointer = Long.valueOf((String) prop.get(filepath));
logger.info("=ExecRunnable.run=>filePointer get from  Properties");
}catch (Exception ex){
logger.error("=ExecRunnable.run=>",ex);
logger.info("=ExecRunnable.run=> error filePointer get from file size");
}
}
if(filePointer ==null){
filePointer = this.logfile.length(); //指针标识从文件的当前长度开始。
logger.info("=ExecRunnable.run=>filePointer get from file size");
}

}

final List<Event> eventList = new ArrayList<Event>();

timedFlushService = Executors.newSingleThreadScheduledExecutor(
new ThreadFactoryBuilder().setNameFormat(
"timedFlushExecService" +
Thread.currentThread().getId() + "-%d").build());
RandomAccessFile randomAccessFile = null;
try {

randomAccessFile= new RandomAccessFile(logfile, "r"); //创建随机读写文件
future = timedFlushService.scheduleWithFixedDelay(new Runnable() {
@Override
public void run() {
try {
synchronized (eventList) {
if(!eventList.isEmpty() && timeout()) {
flushEventBatch(eventList);
}
}
} catch (Exception e) {
logger.error("Exception occured when processing event batch", e);
if(e instanceof InterruptedException) {
Thread.currentThread().interrupt();
}
}
}
},
batchTimeout, batchTimeout, TimeUnit.MILLISECONDS);

while (this.tailing) {
long fileLength = this.logfile.length();
if (fileLength < filePointer) {
randomAccessFile = new RandomAccessFile(logfile, "r");
filePointer = 0l;
}
if (fileLength > filePointer) {
randomAccessFile.seek(filePointer);
String line = randomAccessFile.readLine();
while (line != null) {

//送channal
synchronized (eventList) {
sourceCounter.incrementEventReceivedCount();

String bodyjson = "";
if (!contextIsJson) {
HashMap body = new HashMap();
body.put("context", line.toString());
body.put("filepath", filepath);
body.put("contextType", filepath);
body.put("created", System.currentTimeMillis());
body.put("localHostIp", HostUtils.getLocalHostIp());
body.put("localHostName", HostUtils.getLocalHostName());
body.put("domain", getDomain(filepath));
String context = line.toString();
if(contextIsFlumeLog){
body = ParseFlumeLog(context,body);
}
bodyjson = JSON.toString(body);
if(bodyjson.indexOf("{")>0){
bodyjson.substring(bodyjson.indexOf("{"),bodyjson.length()-1);
}
} else {
bodyjson = line.toString();
}

Event oneEvent = EventBuilder.withBody(bodyjson.getBytes(charset));
eventList.add(oneEvent);
if (eventList.size() >= bufferCount || timeout()) {
flushEventBatch(eventList);
}
}

//读下一行
line = randomAccessFile.readLine();
try {
Long nowfilePointer = randomAccessFile.getFilePointer();
if (!nowfilePointer.equals(filePointer)) {
filePointer = nowfilePointer;
if (System.currentTimeMillis() - timepoint > flushTime) {
timepoint = System.currentTimeMillis();
prop.setProperty(filepath, filePointer.toString());
FileOutputStream fos = new FileOutputStream(fileWriteJson);
if (fos != null) {
prop.store(fos, "Update '" + filepath + "' value");
}
fos.close();

}
}
}catch(Exception ex){
ex.printStackTrace();
}
}

}
Thread.sleep(this.readinterval);
}

synchronized (eventList) {
if(!eventList.isEmpty()) {
flushEventBatch(eventList);
}
}

} catch (Exception e) {
logger.error("Failed while running filpath: " + filepath, e);
if(e instanceof InterruptedException) {
Thread.currentThread().interrupt();
}
} finally {

if(randomAccessFile!=null){
try {
randomAccessFile.close();
} catch (IOException ex) {
logger.error("Failed to close reader for ExecTail source", ex);
}
}

}
logger.info("=run=> flume tail source run restart:"+restart);
if(restart) {
logger.info("=run=> flume tail source run restart time:"+new Date().toString());
logger.info("Restarting in {}ms", restartThrottle);
try {
Thread.sleep(restartThrottle);
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
} else {
logger.info("filepath [" + filepath + "] exited with restart[" + restart+"]");
}
} while(restart);
}

private void flushEventBatch(List<Event> eventList){
channelProcessor.processEventBatch(eventList);
sourceCounter.addToEventAcceptedCount(eventList.size());
eventList.clear();
lastPushToChannel = systemClock.currentTimeMillis();
}

private HashMap ParseFlumeLog(String log,HashMap logMap){
String[] strLogs = log.split("\\|");
logMap.put("className",strLogs[0]);
logMap.put("methodName",strLogs[1]);
logMap.put("level",strLogs[2]);
logMap.put("treeId",strLogs[3]);
logMap.put("requestId",strLogs[4]);
logMap.put("transactionId",strLogs[5]);
return logMap;
}

private boolean timeout(){
return (systemClock.currentTimeMillis() - lastPushToChannel) >= batchTimeout;
}

private static String[] formulateShellCommand(String shell, String command) {
String[] shellArgs = shell.split("\\s+");
String[] result = new String[shellArgs.length + 1];
System.arraycopy(shellArgs, 0, result, 0, shellArgs.length);
result[shellArgs.length] = command;
return result;
}

public int kill() {
logger.info("=kill=> flume tail source kill start time:"+new Date().toString());
this.tailing=false;
synchronized (this.getClass()) {
try {
// Stop the Thread that flushes periodically
if (future != null) {
future.cancel(true);
}

if (timedFlushService != null) {
timedFlushService.shutdown();
while (!timedFlushService.isTerminated()) {
try {
timedFlushService.awaitTermination(500, TimeUnit.MILLISECONDS);
} catch (InterruptedException e) {
logger.debug("Interrupted while waiting for ExecTail executor service "
+ "to stop. Just exiting.");
Thread.currentThread().interrupt();
}
}
}
logger.info("=kill=> flume tail source kill end time:" + new Date().toString());
return Integer.MIN_VALUE;
} catch (Exception ex) {
logger.error("=kill=>", ex);
Thread.currentThread().interrupt();
}
}
logger.info("=kill=> flume tail source kill end time:"+new Date().toString());
return Integer.MIN_VALUE / 2;
}
public void setRestart(boolean restart) {
this.restart = restart;
}
}
private static class StderrReader extends Thread {
private BufferedReader input;
private boolean logStderr;

protected StderrReader(BufferedReader input, boolean logStderr) {
this.input = input;
this.logStderr = logStderr;
}

@Override
public void run() {
try {
int i = 0;
String line = null;
while((line = input.readLine()) != null) {
if(logStderr) {
// There is no need to read 'line' with a charset
// as we do not to propagate it.
// It is in UTF-16 and would be printed in UTF-8 format.
logger.info("StderrLogger[{}] = '{}'", ++i, line);
}
}
} catch (IOException e) {
logger.info("StderrLogger exiting", e);
} finally {
try {
if(input != null) {
input.close();
}
} catch (IOException ex) {
logger.error("Failed to close stderr reader for ExecTail source", ex);
}
}
}
}
}

总结：

这是初步的实现，通过多线程与file i/o 协作，在效率和实时性上，做平衡，实现一版简单的断点续传，以前看似高大上的功能已经被破解一小部分，在讨论中，实际还有结果的收集与去重和正确性验证，加密，等，综合成本考虑，这版简单的，暂行，我们又为flume贡献了一大功能，希望大家一起，将flume做的更完善！