Timeout when running for a while

[feliwir]

When i have DICOM-RST running for a longer time i get timeoutes, e.g.

Jul 08 11:20:01 v220260149660424389 dicom-rst[2205189]: 2026-07-08T09:20:01.839591Z INFO request:studies: Failed to store instance: Timed out sop_instance_uid="1.3.12.2.1107.5.1.4.73318.30000025101706151377800079792" method=POST uri=/aets/SynapseDicomSCP/studies version=HTTP/1.1
Jul 08 11:20:01 v220260149660424389 dicom-rst[2205189]: 2026-07-08T09:20:01.993081Z INFO request:studies: Failed to store instance: Timed out sop_instance_uid="1.3.46.670589.11.29090.5.0.12400.20260528144403561" method=POST uri=/aets/SynapseDicomSCP/studies version=HTTP/1.1
Jul 08 11:20:10 v220260149660424389 dicom-rst[2205189]: 2026-07-08T09:20:10.799304Z INFO request:studies: Failed to store instance: Timed out sop_instance_uid="1.3.12.2.1107.5.2.20.72005.2025102011372126739436784" method=POST uri=/aets/SynapseDicomSCP/studies version=HTTP/1.1
Jul 08 11:20:10 v220260149660424389 dicom-rst[2205189]: 2026-07-08T09:20:10.913713Z INFO request:studies: Failed to store instance: Timed out sop_instance_uid="1.3.12.2.1107.5.1.4.73318.30000025101706151377800079260" method=POST uri=/aets/SynapseDicomSCP/studies version=HTTP/1.1
Jul 08 11:20:11 v220260149660424389 dicom-rst[2205189]: 2026-07-08T09:20:11.481600Z INFO request:studies: Failed to store instance: Timed out sop_instance_uid="1.3.12.2.1107.5.2.18.41630.30000026061117195609900000469" method=POST uri=/aets/SynapseDicomSCP/studies version=HTTP/1.1

Is it possible that the association pools hold on to connections that are not valid anymore?
My setup in this case is "DICOM-RST -> VPN -> PACS". I think the VPN might invalidate some of the connections after a period of time.

Can i disable the association pooling altogether?

[feliwir]

After some more resarching it turns that associations also become unusable after an error event, e.g.

Jul 14 10:11:17 v220260149660424389 dicom-rst[3431326]: 2026-07-14T08:11:17.287373Z ERROR ClientAssociation: Failed to connect: no presentation contexts accepted by the server association_uuid="f384bd77-e1c4-4cb1-bcf1-ccac030e314f" association_id="f384bd77-e1c4-4cb1-bcf1-ccac030e314f"

After this all associations are again in a deadly timeout loop.
You encountered something similar before @nickamzol ?

[feliwir]

Actually i wasn't on the latest commit. The deadloop behaviour is not present anymore. However the STOW-RS request returns no error eventough we get an issue about no valid presentation context. I'll investigate some more