Home > Article > Backend Development > How to solve the problem of concurrent reading and writing of files in PHP
In the case of high concurrency, operating the same file will cause data confusion, so some special processing is required when operating the file. Here are several solutions. Hope it helps everyone.
<span style="font-size: 14px;">/* <br/>* flock(file,lock,block) <br/>* file 必需,规定要锁定或释放的已打开的文件 <br/>* lock 必需。规定要使用哪种锁定类型。 <br/>* block 可选。若设置为 1 或 true,则当进行锁定时阻挡其他进程。 <br/>* lock <br/>* LOCK_SH 要取得共享锁定(读取的程序) <br/>* LOCK_EX 要取得独占锁定(写入的程序) <br/>* LOCK_UN 要释放锁定(无论共享或独占) <br/>* LOCK_NB 如果不希望 flock() 在锁定时堵塞 <br/>/* <br/><br/>// 获取锁if (flock($file,LOCK_EX)) { // 操作文件<br/> fwrite($file,'write more words'); <br/> // 操作完毕后释放锁<br/> flock($file,LOCK_UN); <br/>} else { <br/> //处理错误逻辑 <br/>} <br/>fclose($file); <br/></span>
The flock function seems to often be exclusive under multiple concurrent situations Resources are not released immediately, or not released at all, resulting in deadlock, which causes the server's CPU usage to be very high, and sometimes even causes the server to die completely. Therefore, simply using flock to lock cannot completely solve the problem.
<span style="font-size: 14px;">if($fp = fopen($fileName,'a')) { $startTime = microtime(); do{ $canWrite = flock($fp, LOCK_EX); if(!$canWrite) {<br/> usleep(round(rand(0, 100) * 1000)); // 释放cpu,将cpu资源先让给其他进程<br/> } // 如果未获取到锁,且未超时,则继续获取锁<br/> } while((!$canWrite) && ((microtime() - $startTime) < 1000)); if($canWrite) { fwrite($fp, $dataToSave);<br/> flock($file,LOCK_UN); <br/> } fclose($fp);<br/>}<br/></span>
<span style="font-size: 14px;">$dir_fileopen='tmp';function cfopen($filename,$mode){<br/> global $dir_fileopen;<br/> clearstatcache(); // 创建一个临时文件<br/> do{ $id=uniqid(); $tempfilename=$dir_fileopen.'/'.$id.md5($filename);<br/> } while(file_exists($tempfilename)); // 将要操作的文件内容拷贝到临时文件中<br/> copy($filename,$tempfilename); $fp = fopen($tempfilename, $mode); return $fp ? [$fp, $filename, $id, @filemtime($filename)] : false;<br/>}function cfwrite($fp,$string){<br/> // 将新增内容写入到临时文件中<br/> return fwrite($fp[0], $string);<br/>}function cfclose($fp){<br/> global $dir_fileopen; $success = fclose($fp[0]);<br/> clearstatcache(); $tempfilename = $dir_fileopen.'/'.$fp[2].md5($fp[1]); // 如果要操作的文件在操作期间没有被修改过,则说明没有人操作过该文件,那么将临时文件改名为真正的文件<br/> if((@filemtime($fp[1]) == $fp[3])){<br/> rename($tempfilename,$fp[1]);<br/> }else{ //说明有其它进程在操作目标文件,当前进程被拒绝,删除临时文件<br/> unlink($tempfilename); $success = false;<br/> } return $success;<br/>}$startTime = microtime();do{ $fp=cfopen('lock.txt','a+');<br/> cfwrite($fp,"welcome to beijing.\n"); $success = cfclose($fp, 'on'); if(!$success) {<br/> usleep(round(rand(0, 100) * 1000)); // 释放cpu,将cpu资源先让给其他进程<br/> }<br/>}while(!$success && ((microtime() - $startTime) < 1000)); // 如果为false,说明操作失败,则重新进行一次操作<br/></span>
Create a queue for file operations, then write a script to read the file operation information from the queue in sequence and then perform corresponding operations on the file. , so that only one process is operating the file at a time, which solves the concurrency problem.
The above content is referenced from: http://blog.csdn.net/daiyan_csdn/article/details/51524781
In situations where concurrency is high In this case, operating the same file will cause data confusion, so some special processing needs to be performed when operating the file. Here are some solutions.
<span style="font-size: 14px;">/* <br/>* flock(file,lock,block) <br/>* file 必需,规定要锁定或释放的已打开的文件 <br/>* lock 必需。规定要使用哪种锁定类型。 <br/>* block 可选。若设置为 1 或 true,则当进行锁定时阻挡其他进程。 <br/>* lock <br/>* LOCK_SH 要取得共享锁定(读取的程序) <br/>* LOCK_EX 要取得独占锁定(写入的程序) <br/>* LOCK_UN 要释放锁定(无论共享或独占) <br/>* LOCK_NB 如果不希望 flock() 在锁定时堵塞 <br/>/* <br/><br/>// 获取锁if (flock($file,LOCK_EX)) { // 操作文件<br/> fwrite($file,'write more words'); <br/> // 操作完毕后释放锁<br/> flock($file,LOCK_UN); <br/>} else { <br/> //处理错误逻辑 <br/>} <br/>fclose($file); <br/></span>
The flock function seems to often monopolize resources under multiple concurrent situations. Immediate release, or no release at all, causes deadlock, which causes the server's CPU usage to be very high, and sometimes even causes the server to die completely. Therefore, simply using flock to lock cannot completely solve the problem.
<span style="font-size: 14px;">if($fp = fopen($fileName,'a')) { $startTime = microtime(); do{ $canWrite = flock($fp, LOCK_EX); if(!$canWrite) {<br/> usleep(round(rand(0, 100) * 1000)); // 释放cpu,将cpu资源先让给其他进程<br/> } // 如果未获取到锁,且未超时,则继续获取锁<br/> } while((!$canWrite) && ((microtime() - $startTime) < 1000)); if($canWrite) { fwrite($fp, $dataToSave);<br/> flock($file,LOCK_UN); <br/> } fclose($fp);<br/>}<br/></span>
<span style="font-size: 14px;">$dir_fileopen='tmp';function cfopen($filename,$mode){<br/> global $dir_fileopen;<br/> clearstatcache(); // 创建一个临时文件<br/> do{ $id=uniqid(); $tempfilename=$dir_fileopen.'/'.$id.md5($filename);<br/> } while(file_exists($tempfilename)); // 将要操作的文件内容拷贝到临时文件中<br/> copy($filename,$tempfilename); $fp = fopen($tempfilename, $mode); return $fp ? [$fp, $filename, $id, @filemtime($filename)] : false;<br/>}function cfwrite($fp,$string){<br/> // 将新增内容写入到临时文件中<br/> return fwrite($fp[0], $string);<br/>}function cfclose($fp){<br/> global $dir_fileopen; $success = fclose($fp[0]);<br/> clearstatcache(); $tempfilename = $dir_fileopen.'/'.$fp[2].md5($fp[1]); // 如果要操作的文件在操作期间没有被修改过,则说明没有人操作过该文件,那么将临时文件改名为真正的文件<br/> if((@filemtime($fp[1]) == $fp[3])){<br/> rename($tempfilename,$fp[1]);<br/> }else{ //说明有其它进程在操作目标文件,当前进程被拒绝,删除临时文件<br/> unlink($tempfilename); $success = false;<br/> } return $success;<br/>}$startTime = microtime();do{ $fp=cfopen('lock.txt','a+');<br/> cfwrite($fp,"welcome to beijing.\n"); $success = cfclose($fp, 'on'); if(!$success) {<br/> usleep(round(rand(0, 100) * 1000)); // 释放cpu,将cpu资源先让给其他进程<br/> }<br/>}while(!$success && ((microtime() - $startTime) < 1000)); // 如果为false,说明操作失败,则重新进行一次操作<br/></span>
Create a queue for file operations, then write a script to read the file operation information from the queue in sequence and then perform corresponding operations on the file. , so that only one process is operating the file at a time, which solves the concurrency problem.
Related recommendations:
Solving the problem of concurrent reading and writing file conflicts in PHP
The above is the detailed content of How to solve the problem of concurrent reading and writing of files in PHP. For more information, please follow other related articles on the PHP Chinese website!