- πΊπΈUnited States jrockowitz Brooklyn, NY
Are you using a load balancer?
- Status changed to Postponed: needs info
over 1 year ago 3:31am 24 June 2023 - Status changed to Active
over 1 year ago 5:33pm 26 June 2023 - π¨π¦Canada drupalthings
No, only a single windows 10 dev machine running an Ubuntu 20.04 install. Drupal and mysql installed on the ubuntu vm with the public and private filesystem located on the Ubuntu file system. No load balancing.
- π¨π¦Canada drupalthings
Good Point. If you accept the premise that the archive needs to be closed before it is opened again, you could put in some print statements to the logfile each time the archive is created and opened and a file is added. You would need a webform with at least a single file upload field that accepts image files (probably any large file would do). You could get away with a smaller batch size if you changed the minimum size of the batch to less than 100 (I think it might be hardcoded as such). It's interesting to note in the documentation and comments on zipArchive at https://www.php.net/manual/en/ziparchive.close.php#93322.
Pay attention, that ZipArchive::addFile() only opens file descriptors and does not compress it. And only ZipArchive::close() compress file and it takes quite a lot of time. Be careful with timeouts.
This little program creates some temporary 2MB files and uploads them in two batches of 10 files each. On my system at least, only the last 10 get added if I comment out the close. If I close the file after each batch all 20 get added:
<?php class BatchArchive { private $zip; private $archiveName; public function __construct($archiveName) { $this->archiveName = $archiveName; $this->zip = new ZipArchive(); $flags = (file_exists($this->archiveName)) ? NULL : ZipArchive::CREATE | ZipArchive::OVERWRITE; echo "Opening archive {$this->archiveName} with flags: {$flags}\n"; if ($this->zip->open($this->archiveName, $flags) !== true) { die("Failed to create or open archive: $this->archiveName"); } } public function addBatchToArchive($files) { foreach ($files as $file) { $localName = basename($file); $this->zip->addFile($file, $localName); echo "Added file to archive: $localName\n"; } } public function closeArchive() { //$this->zip->close(); //echo "Archive closed: $this->archiveName\n"; } } $numberOfFiles = 20; $filesPerBatch = $numberOfFiles / 2; // Create temporary files $tempFiles = []; for ($i = 1; $i <= $numberOfFiles; $i++) { $tempFileName = tempnam(sys_get_temp_dir(), 'tempfile' . $i); $fileSize = 2 * 1024 * 1024; // 2MB $randomData = openssl_random_pseudo_bytes($fileSize); file_put_contents($tempFileName, $randomData); $tempFiles[] = $tempFileName; echo "Created temporary file: $tempFileName\n"; } // Divide files into batches $batches = array_chunk($tempFiles, $filesPerBatch); // Create batch objects and add batches to archive $archiveName = 'archive.zip'; if (file_exists($archiveName)) { unlink($archiveName); echo "Deleted existing archive: $archiveName\n"; } foreach ($batches as $batch) { echo "Adding batch to archive\n"; $batchArchive = new BatchArchive($archiveName); $batchArchive->addBatchToArchive($batch); $batchArchive->closeArchive(); } ?>
And the output with the close archive commented out is:
dpostle-tech@drupal-develop:~/projects/temp$ php test9.php Created temporary file: /tmp/tempfile1jeztvO Created temporary file: /tmp/tempfile2OfQlwM Created temporary file: /tmp/tempfile3raKBLO Created temporary file: /tmp/tempfile42DgOhM Created temporary file: /tmp/tempfile5phKybP Created temporary file: /tmp/tempfile6qlSfsP Created temporary file: /tmp/tempfile7mzDsHM Created temporary file: /tmp/tempfile8jTrHKN Created temporary file: /tmp/tempfile9P5qLGM Created temporary file: /tmp/tempfile10H0hH2O Created temporary file: /tmp/tempfile11VKTCdP Created temporary file: /tmp/tempfile12CnPzUL Created temporary file: /tmp/tempfile13Uz0dxN Created temporary file: /tmp/tempfile14BEo6SM Created temporary file: /tmp/tempfile15wyjHFM Created temporary file: /tmp/tempfile16poB9cP Created temporary file: /tmp/tempfile17epq1yM Created temporary file: /tmp/tempfile180irQ0O Created temporary file: /tmp/tempfile19yr5BnO Created temporary file: /tmp/tempfile20SK5v3M Deleted existing archive: archive.zip Adding batch to archive Opening archive archive.zip with flags: 9 Added file to archive: tempfile1jeztvO Added file to archive: tempfile2OfQlwM Added file to archive: tempfile3raKBLO Added file to archive: tempfile42DgOhM Added file to archive: tempfile5phKybP Added file to archive: tempfile6qlSfsP Added file to archive: tempfile7mzDsHM Added file to archive: tempfile8jTrHKN Added file to archive: tempfile9P5qLGM Added file to archive: tempfile10H0hH2O Adding batch to archive Opening archive archive.zip with flags: 9 Added file to archive: tempfile11VKTCdP Added file to archive: tempfile12CnPzUL Added file to archive: tempfile13Uz0dxN Added file to archive: tempfile14BEo6SM Added file to archive: tempfile15wyjHFM Added file to archive: tempfile16poB9cP Added file to archive: tempfile17epq1yM Added file to archive: tempfile180irQ0O Added file to archive: tempfile19yr5BnO Added file to archive: tempfile20SK5v3M dpostle-tech@drupal-develop:~/projects/temp$ unzip -l archive.zip Archive: archive.zip Length Date Time Name --------- ---------- ----- ---- 2097152 2023-06-27 13:58 tempfile11VKTCdP 2097152 2023-06-27 13:58 tempfile12CnPzUL 2097152 2023-06-27 13:58 tempfile13Uz0dxN 2097152 2023-06-27 13:58 tempfile14BEo6SM 2097152 2023-06-27 13:58 tempfile15wyjHFM 2097152 2023-06-27 13:58 tempfile16poB9cP 2097152 2023-06-27 13:58 tempfile17epq1yM 2097152 2023-06-27 13:58 tempfile180irQ0O 2097152 2023-06-27 13:58 tempfile19yr5BnO 2097152 2023-06-27 13:58 tempfile20SK5v3M --------- ------- 20971520 10 files dpostle-tech@drupal-develop:~/projects/temp$
The key point here is even though the first batch was added to the file, it didn't exist and so was created again when adding the second batch
- π¨π¦Canada drupalthings
As for a precise situation for reproducing, I was able to reproduce using a simple test form with one field for uploading files.
- I used the test functionality of the form to create 12 forms,
- I then created a temporary file of 2MB and copied it to the private/default/webforms/test_downloads submission directories for each uploaded file,
- I changed "Batch Export Size" to 10 under Configuration->Advanced webform settings (the min size is not limited to 100, I misread that),
- made sure the tmp directory was clear of any partial archives,
- Used drush to clear the cache and downloaded the results with "Download uploaded files" checked.
There appears to be a timing aspect related to the size of the uploaded files -- I couldn't get it to happen with the tiny uploaded files created by the test form tab without replacing the uploaded files with the larger 2MB versions.
- π¨π¦Canada hargurpreet Kitchener
I have also experienced the same issue while exporting the both submissions and uploaded files. To fix it, I have created this patch which works fine for me. Thanks!
- πΊπΈUnited States fizcs3 Omaha, Nebraska; USA
We did also have an very odd issue downloading a zip archive whereas it would skip submissions. It didn't happen for all webforms, and really can't characterize it any better than it involved choosing:
* Export Format: PDF documents
* Download uploaded files: checked/yesAll I can say is applied the small patch in #8 and it fixed it.
Am confirming the patch applies on Drupal 10.2.6 with webform 6.2.2.
Thank you @hargurpreet - Status changed to Needs review
5 months ago 3:53pm 24 June 2024 - πΊπΈUnited States jrockowitz Brooklyn, NY
I am unsure if the patch is getting to the root cause. It seems that the archive is being closed and needs to be reopened when zipping large files.
@see https://stackoverflow.com/questions/16121885/php-zip-archive-memory-ram-...
I would be more comfortable with a patch that checks if the archive is closed and then reopens it. The current patch is reopening the archive with every file.