If I have a php script that does nothing but the following...
$file = fopen($path, "r");
flock($file, LOCK_SH);
echo fread($file, filesize($path));
flock($file, LOCK_UN);
fclose($file);
...how much overhead is caused by accessing the script from a browser as opposed to simply accessing the actual file? Internally, is the entire file copied to some kind of buffer and then spit out again or is it nearly the same thing?
Yes, it's read into a buffer. It's essentially equivalent to:
$temp = fread($file, filesize($path));
echo $temp;
You can use fpassthru() to send to the client without reading everything into a buffer at once.
fpassthru($file);
I have the code below to output a big file, but it's falling over because PHP's memory use seems to grow and grow as the file is read:
<?php
// various header() calls etc.
$stream = fopen($tarfile,'r');
ob_end_flush();
while (!feof($stream)) {
$buf = fread($stream, 4096);
print $buf;
flush();
unset($buf);
$aa_usage = memory_get_usage(TRUE); // ← this keeps going up!
}
fclose($stream);
I had thought that by the combination of flush and unset the additional memory use would be limited to the 4k buffer, but I'm clearly wrong.
If all you need is to output the content of a file then the right tool to do it is the PHP function readfile(). Replace all the code you posted with:
readfile($tarfile);
As the documentation says:
Note:
readfile() will not present any memory issues, even when sending large files, on its own. If you encounter an out of memory error ensure that output buffering is off with ob_get_level().
You can try load only as much data as you need first, and if you load more data use the function: fseek()
I've got a script which receives large file uploads via a PUT request. These files have some processing done on the fly as they are uploading. Sometimes we can detect that a file is invalid in the first few bytes so we die() with an error message. The only problem is that the client still sends the rest of the data which is a huge waste. Is there a way to shutdown the incoming connection?
Code:
$fp = fopen('php://input', 'rb');
// Do some data checking here
if( <invalid> ) {
fclose($fp);
die('Error');
}
stream_socket_shutdown looked like it might do the job but it has no effect.
Is there any way to do this? Even if I have to write an extension just for this?
You may want to give the following a shot and see if this terminates the connection properly:
$fp = fopen('php://input', 'rb');
// Do some data checking here
if( <invalid> ) {
fclose($fp);
header("Content-Length: 0");
header("Connection: close");
flush();
die('Error');
}
I write an php script that help with limit speed and connections in download files. I used fopen() and fseek() something like this:
$f = fopen($file, 'rb');
if($f){
fseek($f,$start);//$start extracted from $_SERVER['HTTP_RANGE']
while(!feof($f)){
echo fread($f,$speed);//$speed is bytes per second
flush();
ob_flush();
sleep(1);
}
fclose($f);
}
download process may take several hours to complete, is whole file be in memory until end of download? and how I can optimize this?
No, fread uses an internal buffer to stream the data (8KB by default), so only a very small part of the file actually resides in memory.
I have a php script on a server to send files to recipents: they get a unique link and then they can download large files. Sometimes there is a problem with the transfer and the file is corrupted or never finishes. I am wondering if there is a better way to send large files
Code:
$f = fopen(DOWNLOAD_DIR.$database[$_REQUEST['fid']]['filePath'], 'r');
while(!feof($f)){
print fgets($f, 1024);
}
fclose($f);
I have seen functions such as
http_send_file
http_send_data
But I am not sure if they will work.
What is the best way to solve this problem?
Regards
erwing
Chunking files is the fastest / simplest method in PHP, if you can't or don't want to make use of something a bit more professional like cURL, mod-xsendfile on Apache or some dedicated script.
$filename = $filePath.$filename;
$chunksize = 5 * (1024 * 1024); //5 MB (= 5 242 880 bytes) per one chunk of file.
if(file_exists($filename))
{
set_time_limit(300);
$size = intval(sprintf("%u", filesize($filename)));
header('Content-Type: application/octet-stream');
header('Content-Transfer-Encoding: binary');
header('Content-Length: '.$size);
header('Content-Disposition: attachment;filename="'.basename($filename).'"');
if($size > $chunksize)
{
$handle = fopen($filename, 'rb');
while (!feof($handle))
{
print(#fread($handle, $chunksize));
ob_flush();
flush();
}
fclose($handle);
}
else readfile($path);
exit;
}
else echo 'File "'.$filename.'" does not exist!';
Ported from richnetapps.com / NeedBee. Tested on 200 MB files, on which readfile() died, even with maximum allowed memory limit set to 1G, that is five times more than downloaded file size.
BTW: I tested this also on files >2GB, but PHP only managed to write first 2GB of file and then broke the connection. File-related functions (fopen, fread, fseek) uses INT, so you ultimately hit the limit of 2GB. Above mentioned solutions (i.e. mod-xsendfile) seems to be the only option in this case.
EDIT: Make yourself 100% that your file is saved in utf-8. If you omit that, downloaded files will be corrupted. This is, because this solutions uses print to push chunk of a file to a browser.
If you are sending truly large files and worried about the impact this will have, you could use the x-sendfile header.
From the SOQ using-xsendfile-with-apache-php, an howto blog.adaniels.nl : how-i-php-x-sendfile/
Best solution would be to rely on lighty or apache, but if in PHP, I would use PEAR's HTTP_Download (no need to reinvent the wheel etc.), has some nice features, like:
Basic throttling mechanism
Ranges (partial downloads and resuming)
See intro/usage docs.
We've been using this in a couple of projects and it works quite fine so far:
/**
* Copy a file's content to php://output.
*
* #param string $filename
* #return void
*/
protected function _output($filename)
{
$filesize = filesize($filename);
$chunksize = 4096;
if($filesize > $chunksize)
{
$srcStream = fopen($filename, 'rb');
$dstStream = fopen('php://output', 'wb');
$offset = 0;
while(!feof($srcStream)) {
$offset += stream_copy_to_stream($srcStream, $dstStream, $chunksize, $offset);
}
fclose($dstStream);
fclose($srcStream);
}
else
{
// stream_copy_to_stream behaves() strange when filesize > chunksize.
// Seems to never hit the EOF.
// On the other handside file_get_contents() is not scalable.
// Therefore we only use file_get_contents() on small files.
echo file_get_contents($filename);
}
}
For downloading files the easiest way I can think of would be to put the file in a temporary location and give them a unique URL that they can download via regular HTTP.
As part generating these links you could also remove files that were more than X hours old.
Create a symbolic link to the actual file and make the download link point at the symbolic link. Then, when the user clicks on the DL link, they'll get a file download from the real file but named from the symbolic link. It takes milliseconds to create the symbolic link and is better than trying to copy the file to a new name and download from there.
For example:
<?php
// validation code here
$realFile = "Hidden_Zip_File.zip";
$id = "UserID1234";
if ($_COOKIE['authvalid'] == "true") {
$newFile = sprintf("myzipfile_%s.zip", $id); //creates: myzipfile_UserID1234.zip
system(sprintf('ln -s %s %s', $realFile, $newFile), $retval);
if ($retval != 0) {
die("Error getting download file.");
}
$dlLink = "/downloads/hiddenfiles/".$newFile;
}
// rest of code
?>
<a href="<?php echo $dlLink; ?>Download File</a>
That's what I did because Go Daddy kills the script from running after 2 minutes 30 seconds or so....this prevents that problem and hides the actual file.
You can then setup a CRON job to delete the symbolic links at regular intervals....
This whole process will then send the file to the browser and it doesn't matter how long it runs since it's not a script.
When I have done this in the past I've used this:
set_time_limit(0); //Set the execution time to infinite.
header('Content-Type: application/exe'); //This was for a LARGE exe (680MB) so the content type was application/exe
readfile($fileName); //readfile will stream the file.
These 3 lines of code will do all the work of the download readfile() will stream the entire file specified to the client, and be sure to set an infinite time limit else you may be running out of time before the file is finished streaming.
If you are using lighttpd as a webserver, an alternative for secure downloads would be to use ModSecDownload. It needs server configuration but you'll let the webserver handle the download itself instead of the PHP script.
Generating the download URL would look like that (taken from the documentation) and it could of course be only generated for authorized users:
<?php
$secret = "verysecret";
$uri_prefix = "/dl/";
# filename
# please note file name starts with "/"
$f = "/secret-file.txt";
# current timestamp
$t = time();
$t_hex = sprintf("%08x", $t);
$m = md5($secret.$f.$t_hex);
# generate link
printf('%s',
$uri_prefix, $m, $t_hex, $f, $f);
?>
Of course, depending on the size of the files, using readfile() such as proposed by Unkwntech is excellent. And using xsendfile as proposed by garrow is another good idea also supported by Apache.
header("Content-length:".filesize($filename));
header('Content-Type: application/zip'); // ZIP file
header('Content-Type: application/octet-stream');
header('Content-Disposition: attachment; filename="downloadpackage.zip"');
header('Content-Transfer-Encoding: binary');
ob_end_clean();
readfile($filename);
exit();
I'm not sure this is a good idea for large files. If the thread for your download script runs until the user has finished the download, and you're running something like Apache, just 50 or more concurrent downloads could crash your server, because Apache isn't designed to run large numbers of long-running threads at the same time. Of course I might be wrong, if the apache thread somehow terminates and the download sits in a buffer somewhere whilst the download progresses.
I have used the following snippet found in the comments of the php manual entry for readfile:
function _readfileChunked($filename, $retbytes=true) {
$chunksize = 1*(1024*1024); // how many bytes per chunk
$buffer = '';
$cnt =0;
// $handle = fopen($filename, 'rb');
$handle = fopen($filename, 'rb');
if ($handle === false) {
return false;
}
while (!feof($handle)) {
$buffer = fread($handle, $chunksize);
echo $buffer;
ob_flush();
flush();
if ($retbytes) {
$cnt += strlen($buffer);
}
}
$status = fclose($handle);
if ($retbytes && $status) {
return $cnt; // return num. bytes delivered like readfile() does.
}
return $status;
}
I have had same problem,
my problem solved by adding this before starting session
session_cache_limiter('none');
This is tested on files of a size 200+ MB on a server that has 256MB memory limit.
header('Content-Type: application/zip');
header("Content-Disposition: attachment; filename=\"$file_name\"");
set_time_limit(0);
$file = #fopen($filePath, "rb");
while(!feof($file)) {
print(#fread($file, 1024*8));
ob_flush();
flush();
}