Your socket connection to the server was not read from or written to within the timeout period. Idle connections will be closed

[nashid]

I am getting this error once I am making a putObject request.

It's not clear to me what is causing this issue. What is the root cause for this error and what would be the strategy to avoid this?

Secondly, how to make this a retryable so that aws-sdk retries while having this error. I am already setting the error retry which does not seem to have any effect.

com.amazonaws.services.s3.model.AmazonS3Exception: Your socket connection to the 
server was not read from or written to within the timeout period. Idle connections 
will be closed. (Service: Amazon S3; Status Code: 400; Error Code: RequestTimeout

Code:

  val s3ClientConfig = new ClientConfiguration().withMaxErrorRetry(3)
    .withConnectionTimeout(60 * 1000)
    .withSocketTimeout(60 * 1000)

In a similar issue (aws/aws-sdk-js#281), it was mentioned this happens when the provided Content-Length is larger than the number of bytes sent, causing S3 to wait for you to send a specified number of bytes then timeout while waiting.

However, for my case, I am not specifying any content length and I am simply invoking with:

PutObjectRequest(String bucketName, String key, File file)

[spfink]

Are you seeing this consistently or intermittently? Can you provide a code sample for where you are calling PutObjectRequest?

[nashid]

This is an intermittent error which we are observing recently and its making our aws batch jobs error-prone and not reliable.

 val s3ClientConfig = new ClientConfiguration().withMaxErrorRetry(3)
    .withConnectionTimeout(60 * 1000)
    .withSocketTimeout(60 * 1000)

 val amazonS3Client: AmazonS3 = AmazonS3ClientBuilder
    .standard()
    .withRegion("us-east-1")
    .withCredentials(new AWSStaticCredentialsProvider(credentials))
    .withClientConfiguration(s3ClientConfig).build

val objectRequest = new PutObjectRequest(s3config.s3Bucket, destination.path, file)

amazonS3Client.putObject(objectRequest)

[nashid]

@spfink do you need anything else from me on this?

[shorea]

Are these files being modified or appended to while you upload? What's the rate at which you upload files?

[nashid]

@shorea we prepare the files locally before uploading - so these files are not modified/appended while we are uploading.

The rate at which we upload is actually very low - at best ~10 upload per hour.

[shorea]

Given your other issue (#1544) I suggest you dive deep into the networking layer to see if there is something going on. If that all checks out I suggest reaching out to the S3 service team with some request ids (where possible).

[shorea]

Closing issue, feel free to reopen if you need further assistance from the SDK team.

[ngbalk]

The OP's questions Secondly, how to make this a retryable so that aws-sdk retries while having this error. I am already setting the error retry which does not seem to have any effect. I think is still a good one that was not answered - how can we make this socket timeout 400 error code retry-able?
@shorea
cc: @debora-ito