将 Parse 文件从 Gridstore 迁移到 S3

Question

我们将数据（没有文件）迁移到 mLab 和 Heroku。所以旧文件仍在 Parse.

从那时起，添加的任何新文件都会进入 Gridstore，这是 mLab 的默认文件存储。

现在我使用 sashido

将旧的解析文件从 Parse 迁移到 S3 Bucket

文件已迁移，可以使用 Heroku 中的 S3Adapter 访问。

但是 Gridstore 上的文件现在无法访问。如何将它们迁移到同一个 S3 存储桶 并更改 mLab 中的引用？

Answer 1

也许您对我试过的解决方案感兴趣。这不是一个简单的操作，但我使用我的解析服务器配置成功迁移了 3 个数据库。

它基于 PHP 脚本（使用 Parse PHP SDK）运行遍历每个对象，它从 Parse.com 获取文件并设置它（使用任何适配器配置）在您自己的服务器中。

脚本如下所示：

<?php
    ini_set('display_errors', 1);
    ini_set('display_startup_errors', 1);
    error_reporting(E_ALL);
    date_default_timezone_set('America/New_York');
    
    $fileField = $argv[1];
    $class     = $argv[2];
    
    require_once 'vendor/autoload.php';
    use Parse\ParseObject;
    use Parse\ParseQuery;
    use Parse\ParseACL;
    use Parse\ParsePush;
    use Parse\ParseUser;
    use Parse\ParseInstallation;
    use Parse\ParseException;
    use Parse\ParseAnalytics;
    use Parse\ParseFile;
    use Parse\ParseCloud;
    use Parse\ParseClient;
    
    $app_id     = "******";
    $rest_key   = "******";
    $master_key = "******";
    
    ParseClient::initialize($app_id, $rest_key, $master_key);
    ParseClient::setServerURL('http://localhost:1338/', 'parse');
    
    $query = new ParseQuery($class);
    $query->ascending("createdAt"); // it's just my preference
    $query->exists($fileField);
    $query->limit(1);
    
    $count = $query->count();
    for ($i = 0; $i < $count; $i = $i + 1) {
        try {
            $query->skip($i);
            // get Entry
            $entryWithFile = $query->first();
            
            // get file
            $parseFile = $entryWithFile->get($fileField);
            // filename
            $fileName  = $parseFile->getName();

            // if the file is hosted in Parse, do the job, otherwise continue with the next one
            if (strpos($fileName, "tfss-") === false) {
                echo "\nThis is already an internal file, skipping...";
                continue;
            }
            
            $newFileName = str_replace("tfss-", "", $fileName);
            $binaryFile  = file_get_contents($parseFile->getURL());
            
            $newFile = ParseFile::createFromData($binaryFile, $newFileName);
            
            $entryWithFile->set($fileField, $newFile);
            $entryWithFile->save(true);
            
            echo "\nFile saved\n";
        }
        catch (Exception $e) {
            // The conection with mongo or the server could be off for some second, let's retry it ;)
            sleep(10);
            continue;
        }
    }
    
    echo "\n";
    echo "END!";
    
?>

正确设置您的解析 url。

假设您要从 class _User 迁移字段为 imageProfile 的文件，因此请务必传递 $fileField = "imageProfile"; $class = "_User".

运行每个字段的代码 class.

我做了一个并行工作的愚蠢解决方案，它会跳过 for 循环中的步骤，例如：

<?php
    ini_set('display_errors', 1);
    ini_set('display_startup_errors', 1);
    error_reporting(E_ALL);
    date_default_timezone_set('America/New_York');
    
    $index     = $argv[1];
    $of        = $argv[2];
    $fileField = $argv[3];
    $class     = $argv[4];
    
    require_once 'vendor/autoload.php';
    use Parse\ParseObject;
    use Parse\ParseQuery;
    use Parse\ParseACL;
    use Parse\ParsePush;
    use Parse\ParseUser;
    use Parse\ParseInstallation;
    use Parse\ParseException;
    use Parse\ParseAnalytics;
    use Parse\ParseFile;
    use Parse\ParseCloud;
    use Parse\ParseClient;
    
    $app_id     = "********";
    $rest_key   = "********";
    $master_key = "********";
    
    ParseClient::initialize($app_id, $rest_key, $master_key);
    ParseClient::setServerURL('http://localhost:1338/', 'parse');
    
    $query = new ParseQuery($class);
    $query->ascending("createdAt");
    $query->exists($fileField);
    $query->limit(1);
    
    $count = $query->count();
    for ($i = $index; $i < $count; $i = $i + $of) {
        try {
            $query->skip($i);
            // get Entry
            $entryWithFile = $query->first();
            
            // get file
            $parseFile = $entryWithFile->get($fileField);
            // filename
            $fileName  = $parseFile->getName();

            // if the file is hosted in Parse, do the job, otherwise continue with the next one
            if (strpos($fileName, "tfss-") === false) {
                echo "\nThis is already an internal file, skipping...";
                continue;
            }
            
            $newFileName = str_replace("tfss-", "", $fileName);
            $binaryFile  = file_get_contents($parseFile->getURL());
            
            $newFile = ParseFile::createFromData($binaryFile, $newFileName);
            
            $entryWithFile->set($fileField, $newFile);
            $entryWithFile->save(true);
            
            echo "\nFile saved\n";
        }
        catch (Exception $e) {
            // The conection with mongo or the server could be off for some second, let's retry it ;)
            sleep(10);
            continue;
        }
    }
    
    echo "\n";
    echo "END!";
    
?>

所以如果你像以前一样配置$fileField和$class，你可以打开3个线程和运行:

php migrator.php 0 3 "imageProfile" "_User"

php migrator.php 1 3 "imageProfile" "_User"

php migrator.php 2 3 "imageProfile" "_User"

所以你会有循环运行ning 像:

对象 0、3、6

对象 1、4、7

对象 2、5、8

祝你好运，快点！过几天就要关了。

将 Parse 文件从 Gridstore 迁移到 S3

Migrating Parse files from Gridstore to S3

parse-platform