java.io.IOException: Connection reset by peer on asynchronous calls of SQS and SES (Channel closed - SSL handshake failure)

[Jahx]

We are getting IOException "Connection reset by peer" for asynchronous SQS and SES clients which is caused by SSLException: "Inbound closed before receiving peer's close_notify: possible truncation attack?"
Please check these screenshots from Intellij debug for more detail: https://photos.app.goo.gl/LCnmULCjzx54nbde9

Expected Behavior

Connections are closed properly. No SSL handshake or channel was closed exceptions

Current Behavior

Please check these screenshots for more detail: https://photos.app.goo.gl/LCnmULCjzx54nbde9

Steps to Reproduce (for bugs)

Make asynchronous call using SesAsynchClinet or SqsAsyncClient. From time to time we get the exception (Usually in 1 minute delay after actual call).

Context

One of our services is used for sending emails. We get the errors and cannot handle it.

Your Environment

• AWS Java SDK version used: 2.2.0
• JDK version used: java-1.8.0 (8u181-b13)
• Operating System and version: docker amazonlinux:latest

[Aleksandr-Filichkin]

Hi, I also see this issue for SES?
Any updates?

[zoewangg]

Having trouble reproducing the issue. It would be helpful if you could provide a more specific scenario to reproduce the issue, such as the TPS to trigger the issue or how long it should be running until the exception occurs.

[Aleksandr-Filichkin]

@zoewangg,
we have some production code that we cannot share. But we will try to write something simple. We will share with you the test app

[baka2010]

@zoewangg
please take a look at https://github.com/baka2010/ses
Just set EMAIL_ADDRESS_TO and EMAIL_ADDRESS_FROM.
This issue is reproduced when you try to send >100 emails after 1 min timeout.
Stack trace screenshot is in repo.

[Aleksandr-Filichkin]

@zoewangg did you have a chance to try this code from the previous comment?

[zoewangg]

@Aleksandr-Filichkin Thanks for providing the sample code, but unfortunately I still wasn't able to reproduce the issue on my laptop. Could you provide the JDK version, OS name and version?

One thing I noticed in your stack trace https://raw.githubusercontent.com/baka2010/ses/master/stack_trace.png is that the logs were WARN logs. Did you actually get exceptions that were causing the requests to fail? If not, it looks like a similar issue to #900 (comment)

[Jahx]

Hi @zoewangg! Thanks for your help. Since it is duplicate issue for #900 we can close this one.