Using BufferedInputStream causes an upload error

[JavaLionLi]

Describe the bug

Upload files to receive using MultipartFile
MultipartFile.getInputStream return BufferedInputStream
Using BufferedInputStream causes an upload error

Expected Behavior

use bytes upload success

Current Behavior

Reproduction Steps

Possible Solution

No response

Additional Information/Context

No response

AWS Java SDK version used

1.12.341

JDK version used

1.8.0_202

Operating System and version

centos7.6

[debora-ito]

Hi @JavaLionLi can you share a repro code in text format, not screenshot?
We ask for repro code for two main reasons: to better understand what the application is trying to do, and to be able to simply copy+paste the code and run it in our side to reproduce the error.

[JavaLionLi]

Hi @JavaLionLi can you share a repro code in text format, not screenshot? We ask for repro code for two main reasons: to better understand what the application is trying to do, and to be able to simply copy+paste the code and run it in our side to reproduce the error.

Simply put, PutObjectRequest use BufferedInputStream error

[JavaLionLi]

Hi @JavaLionLi can you share a repro code in text format, not screenshot? We ask for repro code for two main reasons: to better understand what the application is trying to do, and to be able to simply copy+paste the code and run it in our side to reproduce the error.

        try {
            // The file must be larger than 200kb
            BufferedInputStream inputStream = new BufferedInputStream(new FileInputStream("D:\test.txt"));
            PutObjectRequest putObjectRequest = new PutObjectRequest("test", path, inputStream, new ObjectMetadata());
            client.putObject(putObjectRequest);
        } catch (Exception e) {
            e.printStackTrace();
        }

[debora-ito]

Can you share a code sample of how the s3 client is being created?

[JavaLionLi]

Can you share a code sample of how the s3 client is being created?

        AwsClientBuilder.EndpointConfiguration endpointConfig =
            new AwsClientBuilder.EndpointConfiguration("http://127.0.0.1:9000", "");
        AWSCredentials credentials = new BasicAWSCredentials("admin", "123456");
        AWSCredentialsProvider credentialsProvider = new AWSStaticCredentialsProvider(credentials);
        ClientConfiguration clientConfig = new ClientConfiguration();
        AmazonS3ClientBuilder build = AmazonS3Client.builder()
            .withEndpointConfiguration(endpointConfig)
            .withClientConfiguration(clientConfig)
            .withCredentials(credentialsProvider)
            .disableChunkedEncoding();
        this.client = build.build();

use minio

[debora-ito]

I can't reproduce the issue. I used the s3 endpoint though, not minio. Maybe it's a minio issue, do you see the same error when using the s3 endpoint directly?

You mentioned Multipart in the description, but the repro code calls a regular non-multipart PutObject, can you share a repro code for Multipart Upload? And if possible, please enable the DEBUG logs and share here when the error occurs. For instructions on how to enable DEBUG logs check Request/Response Summary Logging here: https://docs.aws.amazon.com/sdk-for-java/v1/developer-guide/java-dg-logging.html#sdk-net-logging-request-response

[JavaLionLi]

I can't reproduce the issue. I used the s3 endpoint though, not minio. Maybe it's a minio issue, do you see the same error when using the s3 endpoint directly?

You mentioned Multipart in the description, but the repro code calls a regular non-multipart PutObject, can you share a repro code for Multipart Upload? And if possible, please enable the DEBUG logs and share here when the error occurs. For instructions on how to enable DEBUG logs check Request/Response Summary Logging here: https://docs.aws.amazon.com/sdk-for-java/v1/developer-guide/java-dg-logging.html#sdk-net-logging-request-response

Nothing to do with minio, Using a BufferedInputStream file larger than 500kb will result in an error, The error was reported before it was uploaded


BufferedInputStream uses reset to report errors, which can be found on the network

[JavaLionLi]

No one to deal with it?

[hqbhoho]

I had the same problem in sdk 2.x , So I try use java reflection to get BufferedInputSteam filed in (FileInputStream) and put it . @JavaLionLi
code like:

InputStream inputStream = file.getInputStream();
Field in = FilterInputStream.class.getDeclaredField("in");
in.setAccessible(true);
InputStream inNew = (InputStream)in.get(inputStream);
PutObjectResponse putObjectResponse = s3Client.putObject(putObjectRequest, RequestBody.fromInputStream(inNew, size));