TCP connection not closed even after the association is aborted

[shoibahamed]

The PACS after sending images, does not release the association. Hence the "Network timeout" is triggered.
When the association is timed out and aborted by the scp, the TCP connection remains in the ESTAB state. And it will remain in the ESTAB state until the application is restarted

I see this error on the prod environment and the PACS is STENTOR. However, I could not reproduce on my local

New connections are opened for every association and they remain open and eventually and application hits maximum_associations and then rejects all other associations that are trying to be opened with the error message - "local limit exceeded"

[scaramallion]

Can you post your code, the anonymised output from debug_logger(), details of the production environment and which version of pynetdicom you're using?

[scaramallion]

@shoibahamed it's difficult for me to reproduce this without more information.

[shoibahamed]

I'm sorry for delay @scaramallion . I have been trying yo reproduce this and have not been able to. There are no error messages apart from notification that the association was aborted due to network timeout.
This happens only on 2 sites on production

The socket would just stay open until the application is restarted even after the associations is aborted and state machine showing state transitions

The code is running on docker container on image alpine:3.16
pynetdicom == 2.1.0

So I was able to overcome this issue by getting the socket from dul in the association abort event and manually closing it. Like this

socket_obj = assoc.dul.socket
if hasattr(socket_obj, 'socket'):
    socket_info = socket_obj.socket
    socket_info .shutdown(socket.SHUT_RDWR)

[scaramallion]

No problem, thanks for the additional information.