Results are not delivered to PACS

I am not able to configure DICOM server to connect to my external PACS instance and need help. I am able to receive files from my external instance and the pipeline processes them and renders the test image. Nothing happens after that. I am expecting the process files to be submitted to my External PACs instance.

My DICOM yaml SCU configuration;


the host-ip is the ip of the external PACS service listening to port 1004

The command to invoke the scp service:


The command to invoke scu service:

The IP address is of my CLARA DICOM instance.

Thanks for your interest in Clara Deploy!

I think you have everything almost right, but looking at the scu command it looks like you’re missing an argument for port 104. See documentation:

Try adding the port there and let us know if it helps.

Thanks. I have the IP and port 104 blurred out in my image. I can definitely connect to the Clara DICOM service from my local CLI but it is the SCP service listening on my local 1004 port that is not receiving any files. I assume that the Liver Tumor pipeline (from the docs example) is setup to send processed files to an external PACS service.

Thanks for your question. Yes, the Liver and Tumor reference pipeline by default sends the DICOM instances back to as configured external DICOM device listening port. In your case, AE title MYPACS, port 1004 at . The host-ip, in your case, is the IP address of the workstation where you started the scp with sudo scp ...

In the statement the host-ip is the ip of the external PACS service listening to port 1004, does the host-ip refer to the workstation where you started the DICOM scp? If it is referring to another server with its own DICOM listening port, then the destination needs to be refer to that server’s IP, port, as well as AE Title.

The Clara Dicom Adapter’s log can be examined to see if there are any errors trying to send DICOM instances to external listener. This can be done by using Kubernetes commands on the server where Clara Deploy is installed:
kubectl get pods
(find the pod whose name contains the Dicom Adaptor)
kubectl logs <DICOM Adaptor pod>

Hope this helps and best regards

Thanks mingmelvinq - here is the DICOM Adoptor pod log error - no active connection due to operation time out. I checked with ping to make sure my CLARA server can reach the external PACS instance so they are not connecting for some reason, even though the payloads have been downloaded from the job in prep for the export.

Thank you for providing the logs. It is hard to determine what exactly caused the not active connection.

It could be networking issue, i.e. firewall rule prevent inbound traffic to the specific port on your PACS or outbound traffic rule on the Clara Deploy server, or could the configuration for IP, port and AE Title.

Could you please do a DICOM Echo (from the Clara Deploy host) to the intended PACS to test the connection for DICOM protocol? Here is the Guide on how to do it with DCMTK.

Best regards

Yes. Thanks. Firewall it was. Working now! Much appreciated!!

Great! Thanks for trying out NVIDIA Clara!