php 使用 hdfs 分布式存储海量小文件

php可以通过thrift连接hbase，同样php可以通过thrift读取hadoop资源(HDFS资源)。

准备：

php需要thrift的libary

thrift 安装参见 我上一篇文章

packages:hadoop-0.20.2\src\contrib\thriftfs\gen-php // 搜索hadoop 下载源码 ，在hadoop源码包里

<?php
$GLOBALS['THRIFT_ROOT'] = ROOTPATH . '/lib/thrift';
require_once($GLOBALS['THRIFT_ROOT'].'/Thrift.php');
require_once($GLOBALS['THRIFT_ROOT'].'/transport/TSocket.php');
require_once($GLOBALS['THRIFT_ROOT'].'/transport/TBufferedTransport.php');
require_once($GLOBALS['THRIFT_ROOT'].'/protocol/TBinaryProtocol.php');
require_once($GLOBALS["THRIFT_ROOT"] . "/packages/hadoopfs/ThriftHadoopFileSystem.php");
$hadoop_socket = new TSocket("localhost", 59256);

$hadoop_socket -> setSendTimeout(10000); // Ten seconds
$hadoop_socket -> setRecvTimeout(20000); // Twenty seconds
$hadoop_transport = new TBufferedTransport($hadoop_socket);
$hadoop_protocol = new TBinaryProtocol($hadoop_transport);
$hadoopClient = new ThriftHadoopFileSystemClient($hadoop_protocol);
$hadoop_transport -> open();
try {
// create directory
$dirpathname = new Pathname(array("pathname" => "/user/root/hadoop"));
if($hadoopClient -> exists($dirpathname) == TRUE) {
echo $dirpathname -> pathname . " exists.\n";
} else {
$result = $hadoopClient -> mkdirs($dirpathname);
}
// put file
$filepathname = new Pathname(array("pathname" => $dirpathname -> pathname . "/hello.txt"));
$localfile = fopen("hello.txt", "rb");
$hdfsfile = $hadoopClient -> create($filepathname);
while(true) {
$data = fread($localfile, 1024);
if(strlen($data) == 0)
break;
$hadoopClient -> write($hdfsfile, $data);
}
$hadoopClient -> close($hdfsfile);
fclose($localfile);
// get file
echo "read file:\n";
print_r($filepathname);
$data = "";
$hdfsfile = $hadoopClient -> open($filepathname);
print_r($hdfsfile);
while(true) {
$data = $hadoopClient -> read($hdfsfile, 0, 1024);
if(strlen($data) == 0)
break;
print $data;
}
$hadoopClient -> close($hdfsfile);
echo "listStatus:\n";
$result = $hadoopClient -> listStatus($dirpathname);
print_r($result);
foreach($result as $key => $value) {
if($value -> isdir == "1")
print "dir\t";
else
print "file\t";
print $value -> block_replication . "\t" . $value -> length . "\t" . $value -> modification_time . "\t" . $value -> permission . "\t" . $value -> owner . "\t" . $value -> group . "\t" . $value -> path . "\n";
}
$hadoop_transport -> close();
} catch(Exception $e) {
print_r($e);
}
?>

启动hadoop的thrift

hadoop-0.20.2\src\contrib\thriftfs\scripts\start_thrift_server.sh 59256

problem one:

在系统目录创建文件，而不是在hadoop目录中创建文件

原因：

thrift启动时加载默认的配置文件

解决方法：

修改start_thrift_server.sh文件

TOP=/usr/local/hadoop-0.20.2

CLASSPATH=$CLASSPATH:$TOP/conf

problem two:

java.lang.NullPointerException

at org.apache.hadoop.thriftfs.HadoopThriftServer$HadoopThriftHandler.write(HadoopThriftServer.java:282)

at org.apache.hadoop.thriftfs.api.ThriftHadoopFileSystem$Processor$write.process(Unknown Source)

at org.apache.hadoop.thriftfs.api.ThriftHadoopFileSystem$Processor.process(Unknown Source)

at com.facebook.thrift.server.TThreadPoolServer$WorkerProcess.run(Unknown Source)

at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)

at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)

at java.lang.Thread.run(Thread.java:662)

原因：

java返回的map hash id为long类型，而php(32位)无法存储long类型的数据，导致转换成float数据后丢失精度。

private long nextId = new Random().nextLong();

java返回数据：4207488029786584864

php获取数据：4.2074880297866E+18

java获得php传递数据：4207488029786585088

解决方法：

修改hadoop-0.20.2\src\contrib\thriftfs\if\hadoopfs.thrift文件

修改

struct ThriftHandle {

i64 id

}

为

struct ThriftHandle {

string id

}

重新生成php packages

thrift --gen php hadoopfs.thrift

修改org.apache.hadoop.thriftfs.api.ThriftHandle类

修改

public long id;

为：

public String id;

修改相应的程序

org.apache.hadoop.thriftfs.HadoopThriftServer

修改

long id = insert(out);

ThriftHandle obj = new ThriftHandle(id);

为

long id = insert(out);

String _id = String.valueOf(id);

ThriftHandle obj = new ThriftHandle(_id);

修改相应的程序

重新打包，启动hadoop的thrift：

hadoop-0.20.2\src\contrib\thriftfs\scripts\start_thrift_server.sh 59256

这样php就可以连接并且获取hadoop中的资源了