文章目录

前言

Amazon Simple Storage Service（S3），简单存储服务，是一个公开的云存储服务。Web应用程序开发人员可以使用它存储数字资产，包括图片、视频、音乐和文档。S3提供一个RESTful API以编程方式实现与该服务的交互。目前市面上主流的存储厂商都支持S3协议接口。

本文借鉴风希落https://www.cnblogs.com/jsonq/p/18186340大佬的文章及代码修改而来。

项目采用前后端分离模式：
前端：vue3 + element-plus + axios + spark-md5
后端：Springboot 3X + minio+aws-s3 + redis + mysql + mybatisplus

本文全部代码以上传gitee：https://gitee.com/luzhiyong_erfou/learning-notes/tree/master/aws-s3-upload

一、功能展示

上传功能点

  • 大文件分片上传
  • 文件秒传
  • 断点续传
  • 上传进度

下载功能点

  • 分片下载
  • 暂停下载
  • 下载进度

效果展示

在这里插入图片描述

二、思路流程

上传流程

一个文件的上传，对接后端的请求有三个

  • 点击上传时，请求 <检查文件 md5> 接口，判断文件的状态（已存在、未存在、传输部分）
  • 根据不同的状态，通过 <初始化分片上传地址>，得到该文件的分片地址
  • 前端将分片地址和分片文件一一对应进行上传，直接上传至对象存储
  • 上传完毕，调用 <合并文件> 接口，合并文件，文件数据入库
    在这里插入图片描述

整体步骤：

  • 前端计算文件 md5，并发请求查询此文件的状态
  • 若文件已上传，则后端直接返回上传成功，并返回 url 地址
  • 若文件未上传，则前端请求初始化分片接口，返回上传地址。循环将分片文件和分片地址一一对一应 若文件上传一部分，后端会返回该文件的uploadId （minio中的文件标识）和listParts（已上传的分片索引），前端请求初始化分片接口，后端重新生成上传地址。前端循环将已上传的分片过滤掉，未上传的分片和分片地址一一对应。
  • 前端通过分片地址将分片文件一一上传
  • 上传完毕后，前端调用合并分片接口
  • 后端判断该文件是单片还是分片，单片则不走合并，仅信息入库，分片则先合并，再信息入库。删除 redis 中的文件信息，返回文件地址。

下载流程

整体步骤：

  • 前端计算分片下载的请求次数并设置每次请求的偏移长度
  • 循环调用后端接口
  • 后端判断文件是否缓存并获取文件信息，根据前端传入的便宜长度和分片大小获取文件流返回前端
  • 前端记录每片的blob
  • 根据文件流转成的 blob 下载文件

在这里插入图片描述

三、代码示例

service

import cn.hutool.core.bean.BeanUtil;
import cn.hutool.core.date.DateUtil;
import cn.hutool.core.io.FileUtil;
import cn.hutool.core.util.StrUtil;
import cn.hutool.json.JSONUtil;
import cn.superlu.s3uploadservice.common.R;
import cn.superlu.s3uploadservice.config.FileProperties;
import cn.superlu.s3uploadservice.constant.FileHttpCodeEnum;
import cn.superlu.s3uploadservice.mapper.SysFileUploadMapper;
import cn.superlu.s3uploadservice.model.bo.FileUploadInfo;
import cn.superlu.s3uploadservice.model.entity.SysFileUpload;
import cn.superlu.s3uploadservice.model.vo.BaseFileVo;
import cn.superlu.s3uploadservice.model.vo.UploadUrlsVO;
import cn.superlu.s3uploadservice.service.SysFileUploadService;
import cn.superlu.s3uploadservice.utils.AmazonS3Util;
import cn.superlu.s3uploadservice.utils.MinioUtil;
import cn.superlu.s3uploadservice.utils.RedisUtil;
import com.amazonaws.services.s3.model.S3Object;
import com.amazonaws.services.s3.model.S3ObjectInputStream;
import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
import com.baomidou.mybatisplus.extension.service.impl.ServiceImpl;
import jakarta.servlet.http.HttpServletRequest;
import jakarta.servlet.http.HttpServletResponse;
import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import org.springframework.http.HttpStatus;
import org.springframework.http.ResponseEntity;
import org.springframework.stereotype.Service;

import java.io.BufferedOutputStream;
import java.io.IOException;
import java.nio.charset.StandardCharsets;
import java.time.LocalDateTime;
import java.util.List;
import java.util.concurrent.TimeUnit;

@Service
@Slf4j
@RequiredArgsConstructor
public class SysFileUploadServiceImpl extends ServiceImpl<SysFileUploadMapper, SysFileUpload> implements SysFileUploadService {

    private static final Integer BUFFER_SIZE = 1024 * 64; // 64KB

    private final RedisUtil redisUtil;

    private final MinioUtil minioUtil;

    private final AmazonS3Util amazonS3Util;
    private final FileProperties fileProperties;

    /**
     * 检查文件是否存在
     * @param md5
     * @return
     */
    @Override
    public R<BaseFileVo<FileUploadInfo>> checkFileByMd5(String md5) {
        log.info("查询md5: <{}> 在redis是否存在", md5);
        FileUploadInfo fileUploadInfo = (FileUploadInfo)redisUtil.get(md5);

        if (fileUploadInfo != null) {
            log.info("查询到md5:在redis中存在:{}", JSONUtil.toJsonStr(fileUploadInfo));
            if(fileUploadInfo.getChunkCount()==1){
                return R.ok( BaseFileVo.builder(FileHttpCodeEnum.NOT_UPLOADED, null));
            }else{
                List<Integer> listParts = minioUtil.getListParts(fileUploadInfo.getObject(), fileUploadInfo.getUploadId());
//              List<Integer> listParts = amazonS3Util.getListParts(fileUploadInfo.getObject(), fileUploadInfo.getUploadId());
                fileUploadInfo.setListParts(listParts);
                return R.ok( BaseFileVo.builder(FileHttpCodeEnum.UPLOADING, fileUploadInfo));
            }
        }
        log.info("redis中不存在md5: <{}> 查询mysql是否存在", md5);
        SysFileUpload file = baseMapper.selectOne(new LambdaQueryWrapper<SysFileUpload>().eq(SysFileUpload::getMd5, md5));
        if (file != null) {
            log.info("mysql中存在md5: <{}> 的文件 该文件已上传至minio 秒传直接过", md5);
            FileUploadInfo dbFileInfo = BeanUtil.toBean(file, FileUploadInfo.class);
            return R.ok( BaseFileVo.builder(FileHttpCodeEnum.UPLOAD_SUCCESS, dbFileInfo));
        }
        return R.ok( BaseFileVo.builder(FileHttpCodeEnum.NOT_UPLOADED, null));
    }

    /**
     * 初始化文件分片地址及相关数据
     * @param fileUploadInfo
     * @return
     */
    @Override
    public R<BaseFileVo<UploadUrlsVO>> initMultipartUpload(FileUploadInfo fileUploadInfo) {
        log.info("查询md5: <{}> 在redis是否存在", fileUploadInfo.getMd5());
        FileUploadInfo redisFileUploadInfo = (FileUploadInfo)redisUtil.get(fileUploadInfo.getMd5());
        // 若 redis 中有该 md5 的记录，以 redis 中为主
        String object;
        if (redisFileUploadInfo != null) {
            fileUploadInfo = redisFileUploadInfo;
            object = redisFileUploadInfo.getObject();
        } else {
            String originFileName = fileUploadInfo.getOriginFileName();
            String suffix = FileUtil.extName(originFileName);
            String fileName = FileUtil.mainName(originFileName);
            // 对文件重命名，并以年月日文件夹格式存储
            String nestFile = DateUtil.format(LocalDateTime.now(), "yyyy/MM/dd");
            object = nestFile + "/" + fileName + "_" + fileUploadInfo.getMd5() + "." + suffix;

            fileUploadInfo.setObject(object).setType(suffix);
        }
        UploadUrlsVO urlsVO;
        // 单文件上传
        if (fileUploadInfo.getChunkCount() == 1) {
            log.info("当前分片数量 <{}> 单文件上传", fileUploadInfo.getChunkCount());
//            urlsVO = minioUtil.getUploadObjectUrl(fileUploadInfo.getContentType(), object);
            urlsVO=amazonS3Util.getUploadObjectUrl(fileUploadInfo.getContentType(), object);
        } else {
            // 分片上传
            log.info("当前分片数量 <{}> 分片上传", fileUploadInfo.getChunkCount());
//            urlsVO = minioUtil.initMultiPartUpload(fileUploadInfo, object);
            urlsVO = amazonS3Util.initMultiPartUpload(fileUploadInfo, object);
        }
        fileUploadInfo.setUploadId(urlsVO.getUploadId());
        // 存入 redis （单片存 redis 唯一用处就是可以让单片也入库，因为单片只有一个请求，基本不会出现问题）
        redisUtil.set(fileUploadInfo.getMd5(), fileUploadInfo, fileProperties.getOss().getBreakpointTime(), TimeUnit.DAYS);
        return R.ok(BaseFileVo.builder(FileHttpCodeEnum.SUCCESS, urlsVO));
    }

    /**
     * 合并分片
     * @param md5
     * @return
     */
    @Override
    public R<BaseFileVo<String>> mergeMultipartUpload(String md5) {
        FileUploadInfo redisFileUploadInfo = (FileUploadInfo)redisUtil.get(md5);

        String url = StrUtil.format("{}/{}/{}", fileProperties.getOss().getEndpoint(), fileProperties.getBucketName(), redisFileUploadInfo.getObject());
        SysFileUpload files = BeanUtil.toBean(redisFileUploadInfo, SysFileUpload.class);
        files.setUrl(url)
                .setBucket(fileProperties.getBucketName())
                .setCreateTime(LocalDateTime.now());

        Integer chunkCount = redisFileUploadInfo.getChunkCount();
        // 分片为 1 ，不需要合并，否则合并后看返回的是 true 还是 false
        boolean isSuccess = chunkCount == 1 || minioUtil.mergeMultipartUpload(redisFileUploadInfo.getObject(), redisFileUploadInfo.getUploadId());
//        boolean isSuccess = chunkCount == 1 || amazonS3Util.mergeMultipartUpload(redisFileUploadInfo.getObject(), redisFileUploadInfo.getUploadId());
        if (isSuccess) {
            baseMapper.insert(files);
            redisUtil.del(md5);
            return R.ok(BaseFileVo.builder(FileHttpCodeEnum.SUCCESS, url));
        }
        return R.ok(BaseFileVo.builder(FileHttpCodeEnum.UPLOAD_FILE_FAILED, null));
    }

    /**
     * 分片下载
     * @param id
     * @param request
     * @param response
     * @return
     * @throws IOException
     */
    @Override
    public ResponseEntity<byte[]> downloadMultipartFile(Long id, HttpServletRequest request, HttpServletResponse response) throws IOException {
        // redis 缓存当前文件信息，避免分片下载时频繁查库
        SysFileUpload file = null;
        SysFileUpload redisFile = (SysFileUpload)redisUtil.get(String.valueOf(id));
        if (redisFile == null) {
            SysFileUpload dbFile = baseMapper.selectById(id);
            if (dbFile == null) {
                return null;
            } else {
                file = dbFile;
                redisUtil.set(String.valueOf(id), file, 1, TimeUnit.DAYS);
            }
        } else {
            file = redisFile;
        }

        String range = request.getHeader("Range");
        String fileName = file.getOriginFileName();
        log.info("下载文件的 object <{}>", file.getObject());
        // 获取 bucket 桶中的文件元信息，获取不到会抛出异常
//        StatObjectResponse objectResponse = minioUtil.statObject(file.getObject());
        S3Object s3Object = amazonS3Util.statObject(file.getObject());
        long startByte = 0; // 开始下载位置
//        long fileSize = objectResponse.size();
        long fileSize = s3Object.getObjectMetadata().getContentLength();
        long endByte = fileSize - 1; // 结束下载位置
        log.info("文件总长度：{}，当前 range：{}", fileSize, range);

        BufferedOutputStream os = null; // buffer 写入流
//        GetObjectResponse stream = null; // minio 文件流

        // 存在 range，需要根据前端下载长度进行下载，即分段下载
        // 例如：range=bytes=0-52428800
        if (range != null && range.contains("bytes=") && range.contains("-")) {
            range = range.substring(range.lastIndexOf("=") + 1).trim(); // 0-52428800
            String[] ranges = range.split("-");
            // 判断range的类型
            if (ranges.length == 1) {
                // 类型一：bytes=-2343 后端转换为 0-2343
                if (range.startsWith("-")) endByte = Long.parseLong(ranges[0]);
                // 类型二：bytes=2343- 后端转换为 2343-最后
                if (range.endsWith("-")) startByte = Long.parseLong(ranges[0]);
            } else if (ranges.length == 2) { // 类型三：bytes=22-2343
                startByte = Long.parseLong(ranges[0]);
                endByte = Long.parseLong(ranges[1]);
            }
        }

        // 要下载的长度
        // 确保返回的 contentLength 不会超过文件的实际剩余大小
        long contentLength = Math.min(endByte - startByte + 1, fileSize - startByte);
        // 文件类型
        String contentType = request.getServletContext().getMimeType(fileName);

        // 解决下载文件时文件名乱码问题
        byte[] fileNameBytes = fileName.getBytes(StandardCharsets.UTF_8);
        fileName = new String(fileNameBytes, 0, fileNameBytes.length, StandardCharsets.ISO_8859_1);

        // 响应头设置---------------------------------------------------------------------------------------------
        // 断点续传，获取部分字节内容：
        response.setHeader("Accept-Ranges", "bytes");
        // http状态码要为206：表示获取部分内容,SC_PARTIAL_CONTENT,若部分浏览器不支持，改成 SC_OK
        response.setStatus(HttpServletResponse.SC_PARTIAL_CONTENT);
        response.setContentType(contentType);
//        response.setHeader("Last-Modified", objectResponse.lastModified().toString());
        response.setHeader("Last-Modified", s3Object.getObjectMetadata().getLastModified().toString());
        response.setHeader("Content-Disposition", "attachment;filename=" + fileName);
        response.setHeader("Content-Length", String.valueOf(contentLength));
        // Content-Range，格式为：[要下载的开始位置]-[结束位置]/[文件总大小]
        response.setHeader("Content-Range", "bytes " + startByte + "-" + endByte + "/" + fileSize);
//        response.setHeader("ETag", "\"".concat(objectResponse.etag()).concat("\""));
        response.setHeader("ETag", "\"".concat(s3Object.getObjectMetadata().getETag()).concat("\""));
        response.setContentType("application/octet-stream;charset=UTF-8");

        S3ObjectInputStream objectInputStream=null;
        try {
            // 获取文件流
            String object = s3Object.getKey();
            S3Object currentObject = amazonS3Util.getObject(object, startByte, contentLength);
            objectInputStream = currentObject.getObjectContent();
//            stream = minioUtil.getObject(objectResponse.object(), startByte, contentLength);
            os = new BufferedOutputStream(response.getOutputStream());
            // 将读取的文件写入到 OutputStream
            byte[] bytes = new byte[BUFFER_SIZE];
            long bytesWritten = 0;
            int bytesRead = -1;
            while ((bytesRead = objectInputStream.read(bytes)) != -1) {
//            while ((bytesRead = stream.read(bytes)) != -1) {
                if (bytesWritten + bytesRead >= contentLength) {
                    os.write(bytes, 0, (int)(contentLength - bytesWritten));
                    break;
                } else {
                    os.write(bytes, 0, bytesRead);
                    bytesWritten += bytesRead;
                }
            }
            os.flush();
            response.flushBuffer();
            // 返回对应http状态
            return new ResponseEntity<>(bytes, HttpStatus.OK);
        } catch (Exception e) {
            e.printStackTrace();
        } finally {
            if (os != null) os.close();
//            if (stream != null) stream.close();
            if (objectInputStream != null) objectInputStream.close();
        }
        return null;
    }

    @Override
    public R<List<SysFileUpload>> getFileList() {
        List<SysFileUpload> filesList = this.list();
        return R.ok(filesList);
    }

}

AmazonS3Util


import cn.hutool.core.util.IdUtil;
import cn.superlu.s3uploadservice.config.FileProperties;
import cn.superlu.s3uploadservice.constant.FileHttpCodeEnum;
import cn.superlu.s3uploadservice.model.bo.FileUploadInfo;
import cn.superlu.s3uploadservice.model.vo.UploadUrlsVO;
import com.amazonaws.ClientConfiguration;
import com.amazonaws.HttpMethod;
import com.amazonaws.auth.AWSCredentials;
import com.amazonaws.auth.AWSCredentialsProvider;
import com.amazonaws.auth.AWSStaticCredentialsProvider;
import com.amazonaws.auth.BasicAWSCredentials;
import com.amazonaws.client.builder.AwsClientBuilder;
import com.amazonaws.services.s3.AmazonS3;
import com.amazonaws.services.s3.AmazonS3ClientBuilder;
import com.amazonaws.services.s3.model.*;
import com.google.common.collect.HashMultimap;
import io.minio.GetObjectArgs;
import io.minio.GetObjectResponse;
import io.minio.StatObjectArgs;
import io.minio.StatObjectResponse;
import jakarta.annotation.PostConstruct;
import jakarta.annotation.Resource;
import lombok.SneakyThrows;
import lombok.extern.slf4j.Slf4j;
import org.springframework.stereotype.Component;

import java.net.URL;
import java.util.*;
import java.util.stream.Collectors;

@Slf4j
@Component
public class AmazonS3Util {

    @Resource
    private FileProperties fileProperties;

    private AmazonS3 amazonS3;

    // spring自动注入会失败
    @PostConstruct
    public void init() {
        ClientConfiguration clientConfiguration = new ClientConfiguration();
        clientConfiguration.setMaxConnections(100);
        AwsClientBuilder.EndpointConfiguration endpointConfiguration = new AwsClientBuilder.EndpointConfiguration(
                fileProperties.getOss().getEndpoint(), fileProperties.getOss().getRegion());
        AWSCredentials awsCredentials = new BasicAWSCredentials(fileProperties.getOss().getAccessKey(),
                fileProperties.getOss().getSecretKey());
        AWSCredentialsProvider awsCredentialsProvider = new AWSStaticCredentialsProvider(awsCredentials);
        this.amazonS3 = AmazonS3ClientBuilder.standard()
                .withEndpointConfiguration(endpointConfiguration)
                .withClientConfiguration(clientConfiguration)
                .withCredentials(awsCredentialsProvider)
                .disableChunkedEncoding()
                .withPathStyleAccessEnabled(true)
                .build();
    }

    /**
     * 获取 Minio 中已经上传的分片文件
     * @param object 文件名称
     * @param uploadId 上传的文件id（由 minio 生成）
     * @return List<Integer>
     */
    @SneakyThrows
    public List<Integer> getListParts(String object, String uploadId) {
        ListPartsRequest listPartsRequest = new ListPartsRequest( fileProperties.getBucketName(), object, uploadId);
        PartListing listParts = amazonS3.listParts(listPartsRequest);
        return listParts.getParts().stream().map(PartSummary::getPartNumber).collect(Collectors.toList());
    }


    /**
     * 单文件签名上传
     * @param object 文件名称（uuid 格式）
     * @return UploadUrlsVO
     */
    public UploadUrlsVO getUploadObjectUrl(String contentType, String object) {
        try {
            log.info("<{}> 开始单文件上传<>", object);
            UploadUrlsVO urlsVO = new UploadUrlsVO();
            List<String> urlList = new ArrayList<>();
            // 主要是针对图片，若需要通过浏览器直接查看，而不是下载，需要指定对应的 content-type
            HashMultimap<String, String> headers = HashMultimap.create();
            if (contentType == null || contentType.equals("")) {
                contentType = "application/octet-stream";
            }
            headers.put("Content-Type", contentType);

            String uploadId = IdUtil.simpleUUID();
            Map<String, String> reqParams = new HashMap<>();
            reqParams.put("uploadId", uploadId);
            //生成预签名的 URL
            GeneratePresignedUrlRequest generatePresignedUrlRequest = new GeneratePresignedUrlRequest(
                    fileProperties.getBucketName(),
                    object, HttpMethod.PUT);
            generatePresignedUrlRequest.addRequestParameter("uploadId", uploadId);
            URL url = amazonS3.generatePresignedUrl(generatePresignedUrlRequest);
            urlList.add(url.toString());
            urlsVO.setUploadId(uploadId).setUrls(urlList);
            return urlsVO;
        } catch (Exception e) {
            log.error("单文件上传失败: {}", e.getMessage());
            throw new RuntimeException(FileHttpCodeEnum.UPLOAD_FILE_FAILED.getMsg());
        }
    }

    /**
     * 初始化分片上传
     * @param fileUploadInfo 前端传入的文件信息
     * @param object object
     * @return UploadUrlsVO
     */
    public UploadUrlsVO initMultiPartUpload(FileUploadInfo fileUploadInfo, String object) {
        Integer chunkCount = fileUploadInfo.getChunkCount();
        String contentType = fileUploadInfo.getContentType();
        String uploadId = fileUploadInfo.getUploadId();

        log.info("文件<{}> - 分片<{}> 初始化分片上传数据 请求头 {}", object, chunkCount, contentType);
        UploadUrlsVO urlsVO = new UploadUrlsVO();
        try {
            // 如果初始化时有 uploadId，说明是断点续传，不能重新生成 uploadId
            if (uploadId == null || uploadId.equals("")) {
                // 第一步，初始化，声明下面将有一个 Multipart Upload
                // 设置文件类型
                ObjectMetadata metadata = new ObjectMetadata();
                metadata.setContentType(contentType);
                InitiateMultipartUploadRequest initRequest = new InitiateMultipartUploadRequest(fileProperties.getBucketName(),
                        object, metadata);
                uploadId = amazonS3.initiateMultipartUpload(initRequest).getUploadId();
                log.info("没有uploadId，生成新的{}",uploadId);
            }
            urlsVO.setUploadId(uploadId);

            List<String> partList = new ArrayList<>();
            for (int i = 1; i <= chunkCount; i++) {
                //生成预签名的 URL
                //设置过期时间，例如 1 小时后
                Date expiration = new Date(System.currentTimeMillis() + 3600 * 1000);
                GeneratePresignedUrlRequest generatePresignedUrlRequest =
                        new GeneratePresignedUrlRequest(fileProperties.getBucketName(), object,HttpMethod.PUT)
                                .withExpiration(expiration);
                generatePresignedUrlRequest.addRequestParameter("uploadId", uploadId);
                generatePresignedUrlRequest.addRequestParameter("partNumber", String.valueOf(i));
                URL url = amazonS3.generatePresignedUrl(generatePresignedUrlRequest);
                partList.add(url.toString());
            }
            log.info("文件初始化分片成功");
            urlsVO.setUrls(partList);
            return urlsVO;
        } catch (Exception e) {
            log.error("初始化分片上传失败: {}", e.getMessage());
            // 返回 文件上传失败
            throw new RuntimeException(FileHttpCodeEnum.UPLOAD_FILE_FAILED.getMsg());
        }
    }

    /**
     * 合并文件
     * @param object object
     * @param uploadId uploadUd
     */
    @SneakyThrows
    public boolean mergeMultipartUpload(String object, String uploadId) {
        log.info("通过 <{}-{}-{}> 合并<分片上传>数据", object, uploadId, fileProperties.getBucketName());
        //构建查询parts条件
        ListPartsRequest listPartsRequest = new ListPartsRequest(
                fileProperties.getBucketName(),
                object,
                uploadId);
        listPartsRequest.setMaxParts(1000);
        listPartsRequest.setPartNumberMarker(0);
        //请求查询
        PartListing partList=amazonS3.listParts(listPartsRequest);
        List<PartSummary> parts = partList.getParts();
        if (parts==null|| parts.isEmpty()) {
            // 已上传分块数量与记录中的数量不对应，不能合并分块
            throw new RuntimeException("分片缺失，请重新上传");
        }
        // 合并分片
        CompleteMultipartUploadRequest compRequest = new CompleteMultipartUploadRequest(
                fileProperties.getBucketName(),
                object,
                uploadId,
                parts.stream().map(partSummary -> new PartETag(partSummary.getPartNumber(), partSummary.getETag())).collect(Collectors.toList()));
        amazonS3.completeMultipartUpload(compRequest);
        return true;
    }


    /**
     * 获取文件内容和元信息，该文件不存在会抛异常
     * @param object object
     * @return StatObjectResponse
     */
    @SneakyThrows
    public S3Object statObject(String object) {
        return amazonS3.getObject(fileProperties.getBucketName(), object);
    }

    @SneakyThrows
    public S3Object getObject(String object, Long offset, Long contentLength) {
        GetObjectRequest request = new GetObjectRequest(fileProperties.getBucketName(), object);
        request.setRange(offset, offset + contentLength - 1);  // 设置偏移量和长度
        return amazonS3.getObject(request);
    }




}

minioUtil

import cn.hutool.core.util.IdUtil;
import cn.superlu.s3uploadservice.config.CustomMinioClient;
import cn.superlu.s3uploadservice.config.FileProperties;
import cn.superlu.s3uploadservice.constant.FileHttpCodeEnum;
import cn.superlu.s3uploadservice.model.bo.FileUploadInfo;
import cn.superlu.s3uploadservice.model.vo.UploadUrlsVO;
import com.google.common.collect.HashMultimap;
import io.minio.*;
import io.minio.http.Method;
import io.minio.messages.Part;
import jakarta.annotation.PostConstruct;
import jakarta.annotation.Resource;
import lombok.SneakyThrows;
import lombok.extern.slf4j.Slf4j;
import org.springframework.stereotype.Component;

import java.util.*;
import java.util.concurrent.TimeUnit;
import java.util.stream.Collectors;

@Slf4j
@Component
public class MinioUtil {

    private CustomMinioClient customMinioClient;

    @Resource
    private FileProperties fileProperties;

    // spring自动注入会失败
    @PostConstruct
    public void init() {
        MinioAsyncClient minioClient = MinioAsyncClient.builder()
                .endpoint(fileProperties.getOss().getEndpoint())
                .credentials(fileProperties.getOss().getAccessKey(), fileProperties.getOss().getSecretKey())
                .build();
        customMinioClient = new CustomMinioClient(minioClient);
    }

    /**
     * 获取 Minio 中已经上传的分片文件
     * @param object 文件名称
     * @param uploadId 上传的文件id（由 minio 生成）
     * @return List<Integer>
     */
    @SneakyThrows
    public List<Integer> getListParts(String object, String uploadId) {
        ListPartsResponse partResult = customMinioClient.listMultipart(fileProperties.getBucketName(), null, object, 1000, 0, uploadId, null, null);
        return partResult.result().partList().stream()
                .map(Part::partNumber)
                .collect(Collectors.toList());
    }

    /**
     * 单文件签名上传
     * @param object 文件名称（uuid 格式）
     * @return UploadUrlsVO
     */
    public UploadUrlsVO getUploadObjectUrl(String contentType, String object) {
        try {
            log.info("<{}> 开始单文件上传<minio>", object);
            UploadUrlsVO urlsVO = new UploadUrlsVO();
            List<String> urlList = new ArrayList<>();
            // 主要是针对图片，若需要通过浏览器直接查看，而不是下载，需要指定对应的 content-type
            HashMultimap<String, String> headers = HashMultimap.create();
            if (contentType == null || contentType.equals("")) {
                contentType = "application/octet-stream";
            }
            headers.put("Content-Type", contentType);

            String uploadId = IdUtil.simpleUUID();
            Map<String, String> reqParams = new HashMap<>();
            reqParams.put("uploadId", uploadId);
            String url = customMinioClient.getPresignedObjectUrl(GetPresignedObjectUrlArgs.builder()
                    .method(Method.PUT)
                    .bucket(fileProperties.getBucketName())
                    .object(object)
                    .extraHeaders(headers)
                    .extraQueryParams(reqParams)
                    .expiry(fileProperties.getOss().getExpiry(), TimeUnit.DAYS)
                    .build());
            urlList.add(url);
            urlsVO.setUploadId(uploadId).setUrls(urlList);
            return urlsVO;
        } catch (Exception e) {
            log.error("单文件上传失败: {}", e.getMessage());
            throw new RuntimeException(FileHttpCodeEnum.UPLOAD_FILE_FAILED.getMsg());
        }
    }

    /**
     * 初始化分片上传
     * @param fileUploadInfo 前端传入的文件信息
     * @param object object
     * @return UploadUrlsVO
     */
    public UploadUrlsVO initMultiPartUpload(FileUploadInfo fileUploadInfo, String object) {
        Integer chunkCount = fileUploadInfo.getChunkCount();
        String contentType = fileUploadInfo.getContentType();
        String uploadId = fileUploadInfo.getUploadId();

        log.info("文件<{}> - 分片<{}> 初始化分片上传数据 请求头 {}", object, chunkCount, contentType);
        UploadUrlsVO urlsVO = new UploadUrlsVO();
        try {
            HashMultimap<String, String> headers = HashMultimap.create();
            if (contentType == null || contentType.equals("")) {
                contentType = "application/octet-stream";
            }
            headers.put("Content-Type", contentType);

            // 如果初始化时有 uploadId，说明是断点续传，不能重新生成 uploadId
            if (fileUploadInfo.getUploadId() == null || fileUploadInfo.getUploadId().equals("")) {
                uploadId = customMinioClient.initMultiPartUpload(fileProperties.getBucketName(), null, object, headers, null);
            }
            urlsVO.setUploadId(uploadId);

            List<String> partList = new ArrayList<>();
            Map<String, String> reqParams = new HashMap<>();
            reqParams.put("uploadId", uploadId);
            for (int i = 1; i <= chunkCount; i++) {
                reqParams.put("partNumber", String.valueOf(i));
                String uploadUrl = customMinioClient.getPresignedObjectUrl(GetPresignedObjectUrlArgs.builder()
                        .method(Method.PUT)
                        .bucket(fileProperties.getBucketName())
                        .object(object)
                        .expiry(1, TimeUnit.DAYS)
                        .extraQueryParams(reqParams)
                        .build());
                partList.add(uploadUrl);
            }

            log.info("文件初始化分片成功");
            urlsVO.setUrls(partList);
            return urlsVO;
        } catch (Exception e) {
            log.error("初始化分片上传失败: {}", e.getMessage());
            // 返回 文件上传失败
            throw new RuntimeException(FileHttpCodeEnum.UPLOAD_FILE_FAILED.getMsg());
        }
    }

    /**
     * 合并文件
     * @param object object
     * @param uploadId uploadUd
     */
    @SneakyThrows
    public boolean mergeMultipartUpload(String object, String uploadId) {
        log.info("通过 <{}-{}-{}> 合并<分片上传>数据", object, uploadId, fileProperties.getBucketName());
        //目前仅做了最大1000分片
        Part[] parts = new Part[1000];
        // 查询上传后的分片数据
        ListPartsResponse partResult = customMinioClient.listMultipart(fileProperties.getBucketName(), null, object, 1000, 0, uploadId, null, null);
        int partNumber = 1;
        for (Part part : partResult.result().partList()) {
            parts[partNumber - 1] = new Part(partNumber, part.etag());
            partNumber++;
        }
        // 合并分片
        customMinioClient.mergeMultipartUpload(fileProperties.getBucketName(), null, object, uploadId, parts, null, null);
        return true;
    }

    /**
     * 获取文件内容和元信息，该文件不存在会抛异常
     * @param object object
     * @return StatObjectResponse
     */
    @SneakyThrows
    public StatObjectResponse statObject(String object) {
        return customMinioClient.statObject(StatObjectArgs.builder()
                .bucket(fileProperties.getBucketName())
                .object(object)
                .build())
                .get();
    }

    @SneakyThrows
    public GetObjectResponse getObject(String object, Long offset, Long contentLength) {
        return customMinioClient.getObject(GetObjectArgs.builder()
                .bucket(fileProperties.getBucketName())
                .object(object)
                .offset(offset)
                .length(contentLength)
                .build())
                .get();
    }

}

四、疑问

我在全部使用aws-s3上传时出现一个问题至今没有办法解决。只能在查询分片的时候用minio的包进行。

分片后调用amazonS3.listParts()一直超时。

这个问题我在
https://gitee.com/Gary2016/minio-upload/issues/I8H8GM
也看到有人跟我有相同的问题

有解决的朋友麻烦评论区告知下方法。

# aws # 云计算
Logo
华为开发者空间

华为开发者空间，是为全球开发者打造的专属开发空间，汇聚了华为优质开发资源及工具，致力于让每一位开发者拥有一台云主机，基于华为根生态开发、创新。

更多推荐

cover

GaussDB性能调优

avatar 华为开发者空间
cover

领取云主机，带你基于PyTorch构建高效手写体识别系统

avatar 华为开发者空间
cover

华为“云上先锋”实践赛·AI赛道现场会活动成功举办

avatar 华为开发者空间