AWS S3 多部分上传时,如果在上传过程中出现错误或中断,就需要使用断点续传来继续上传。但是有时候会出现无法恢复上传的情况,如上传时断电或者程序异常退出等导致已上传部分数据丢失,就无法继续上传。这种情况下需要重新开始上传。以下是一个示例,演示如何使用AWS SDK进行多部分上传的实现:
import boto3
s3 = boto3.resource('s3')
bucket = s3.Bucket('my-bucket-name')
# Set the desired multipart chunk size
chunk_size = 8 * 1024 * 1024
# Create a new multipart upload and obtain the upload id
mpu = bucket.initiate_multipart_upload(Key='my-large-file.mp4')
upload_id = mpu.id
# Open the file for reading in binary mode
file = open('my-large-file.mp4', 'rb')
try:
# Read the first chunk
data = file.read(chunk_size)
part_number = 1
# Loop over the chunks and upload them
while data:
# Create a new part and upload it
part = mpu.Part(part_number)
response = part.upload(Body=data)
# Print the response
print('Part #{0} uploaded: {1}'.format(part_number, response['ETag']))
# Read the next chunk
data = file.read(chunk_size)
part_number += 1
except Exception as e:
# Abort the upload on error
mpu.abort()
print('Error occurred during upload: {0}'.format(e))
else:
# Complete the upload when done
mpu.complete()
print('Upload completed successfully.')
finally:
# Close the file handle
file.close()
要实现断点续传,需要记录已上传分块的 ETag 和 part number。当出现错误时,可以通过调用 bucket.MultipartUpload('my-large-file.mp4').abort()来终止当前上传任务并删除已上传