RecordFormat format = new DelimitedRecordFormat().withFieldDelimiter(" ");
// sync the filesystem after every 1k tuples
SyncPolicy syncPolicy = new CountSyncPolicy(1000);
// FileRotationPolicy rotationPolicy = new FileSizeRotationPolicy(1.0f, FileSizeRotationPolicy.Units.KB);
/** rotate file with Date,every month create a new file
* format:yyyymm.txt
*/
FileRotationPolicy rotationPolicy = new CountStrRotationPolicy();
FileNameFormat fileNameFormat = new TimesFileNameFormat().withPath("/test/");
RotationAction action = new NewFileAction();
HdfsBolt bolt = new HdfsBolt()
.withFsUrl("hdfs://127.0.0.1:9000")
.withFileNameFormat(fileNameFormat)
.withRecordFormat(format)
.withRotationPolicy(rotationPolicy)
.withSyncPolicy(syncPolicy)
.addRotationAction(action);
然后分别来看各个策略的类。
FileRotationPolicy
import org.apache.storm.hdfs.bolt.rotation.FileRotationPolicy;
import org.apache.storm.tuple.Tuple;
import java.text.SimpleDateFormat;
import java.util.Date;
/**
* 计数以改变Hdfs写入文件的位置，当写入10次的时候，则更改写入文件，更改名字取决于 “TimesFileNameFormat”
* 这个类是线程安全
*/
public class CountStrRotationPolicy implements FileRotationPolicy {
private SimpleDateFormat df = new SimpleDateFormat("yyyyMM");
private String date = null;
private int count = 0;
public CountStrRotationPolicy(){
this.date = df.format(new Date());
// this.date = df.format(new Date());
}
/**
* Called for every tuple the HdfsBolt executes.
*
* @param tuple The tuple executed.
* @param offset current offset of file being written
* @return true if a file rotation should be performed
*/
@Override
public boolean mark(Tuple tuple, long offset) {
count ++;
if(count == 10) {
System.out.print("num :" +count + " ");
count = 0;
return true;
}
else {
return false;
}
}
/**
* Called after the HdfsBolt rotates a file.
*/
@Override
public void reset() {
}
@Override
public FileRotationPolicy copy() {
return new CountStrRotationPolicy();
}
}
FileNameFormat
import org.apache.storm.hdfs.bolt.format.FileNameFormat;
import org.apache.storm.task.TopologyContext;
import java.util.Map;
/**
* 决定重新写入文件时候的名字
* 这里会返回是第几次转换写入文件，将这个第几次做为文件名
*/
public class TimesFileNameFormat implements FileNameFormat {
//默认路径
private String path = "/storm";
//默认后缀
private String extension = ".txt";
private Long times = new Long(0);
public TimesFileNameFormat withPath(String path){
this.path = path;
return this;
}
@Override
public void prepare(Map conf, TopologyContext topologyContext) {
}
@Override
public String getName(long rotation, long timeStamp) {
times ++ ;
//返回文件名，文件名为更换写入文件次数
return times.toString() + this.extension;
}
public String getPath(){
return this.path;
}
}
RotationAction
import org.apache.hadoop.fs.FileContext;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.storm.hdfs.common.rotation.RotationAction;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.io.IOException;
import java.net.URI;
/**
当转换写入文件时候调用的 hook ，这里仅写入日志。
*/
public class NewFileAction implements RotationAction {
private static final Logger LOG = LoggerFactory.getLogger(NewFileAction.class);
@Override
public void execute(FileSystem fileSystem, Path filePath) throws IOException {
LOG.info("Hdfs change the written file！！");
return;
}
}
好久不来iteye,今天又来看看，哈哈,今天碰到在编码时，反射中会抛出
Illegal overloaded getter method with ambiguous type for propert这么个东东，从字面意思看，是反射在获取getter时迷惑了，然后回想起java在boolean值在生成getter时，分别有is和getter，也许我们的反射对象中就有is开头的方法迷惑了jdk，
泛型
在Java SE 1.5之前，没有泛型的情况的下，通过对类型Object的引用来实现参数的“任意化”，任意化的缺点就是要实行强制转换，这种强制转换可能会带来不安全的隐患
泛型的特点：消除强制转换 确保类型安全 向后兼容
简单泛型的定义：
泛型：就是在类中将其模糊化，在创建对象的时候再具体定义
class fan
安装lua_nginx_module 模块
lua_nginx_module 可以一步步的安装，也可以直接用淘宝的OpenResty
Centos和debian的安装就简单了。。
这里说下freebsd的安装：
fetch http://www.lua.org/ftp/lua-5.1.4.tar.gz
tar zxvf lua-5.1.4.tar.gz
cd lua-5.1.4
ma
public class IsAccendListRecursive {
/*递归判断数组是否升序
* if a Integer array is ascending,return true
* use recursion
*/
public static void main(String[] args){
IsAccendListRecursiv
function init() {
var tt = document.getElementById("tt");
var childNodes = tt.childNodes[0].childNodes;
var level = 0;
for (var i = 0; i < childNodes.length; i++) {
在安装hadoop时，执行JPS出现下面错误
[slave16][email protected]:/tmp/hsperfdata_hdfs# jps
Error occurred during initialization of VM
java.lang.Error: Properties init: Could not determine current working
