Server does not close and re-connect on no keep alive ack issue.

[pstreef]

Using the server setup found here and using rsocket-core 1.1.3 my server behaves as expected. If for some reason no acks get delivered during the keep-alive timeout it will restart the connection. In 1.1.4 this no longer works.

Expected Behavior

After:

io.rsocket.exceptions.ConnectionErrorException: No keep-alive acks for x ms

I expect that the connection is closed and my client can reconnect.

Actual Behavior

In 1.1.4 the client RSocketRequester does not properly reset the connection and the next call fails with the same ConnectionErrorException

Steps to Reproduce

I've created a small reproducing example. The test in this projects runs green using rsocket-core 1.1.3, but fails when using 1.1.4

Possible Solution

I believe this is regression from this change.

Your Environment

See sample project

• RSocket version(s) used: 1.1.4 (and 1.1.3)
• Other relevant libraries versions (eg. netty, ...): spring-boot-starter-rsocket 3.2.0
• Platform (eg. JVM version (javar -version) or Node version (node --version)): zulu 17.0.8
• OS and version (eg uname -a): macos 14.2 (23C64)

[david-dvconsulting]

Hi @rstoyanchev, any updates on the planning/status of this bug ?
We're seeing this a lot in our production environment at the moment.

[kevinat]

This bug is serious, it cause we CANNOT recover from temporary network problems, the retry mechanism become useless, the only way to recover is to restart the application, which is woeful.

[kevinat]

Any workaround would be appreciated. @rstoyanchev