search

mercure-imaging / mercure

mercure DICOM Orchestrator
https://mercure-imaging.org
MIT License
65 stars 31 forks source link

Receiver reset problem #58

Closed aarontovars closed 11 months ago

aarontovars commented 11 months ago

Describe the bug

When several DICOMs are sent to mercure from an external machine (not within the VM) using the dcmsend command, the receiver is reset with the following error in the Receiver Logs:

Oct 14 16:45:32 vagrant receiver.sh[13431]: E: 0006:031d TCP I/O Error (Connection reset by peer) occurred in routine: sendAssociationACTCP

Moreover, if I attemp to send a new case after the reset, the dcmsend command execution time will last almost 1 minute due to Requesting association lasting almost a minute.

This behaviour is not seen everytime a file is sent from an external machine, but after several succesfull dcmsend executions. After that, every dcmsend will last almost a minute, compared to only msec in a normal situation.

Also, if I repeat these repetitive dcmsend commands on the mercure's VM, all of them execute as expected at any moment.

tblock79 commented 11 months ago

This error message is created by the storescp tool from the DCMTK package. mercure is using the storescp tool to receive DICOM series. Therefore, this issue is out of mercure's scope. You could check with the OFFIS DCMTK team regarding the error message. The storescp tool is widely used and pretty stable, so I assume that it might have to do with your VM setup (maybe a firewall or CPU load issue that intercepts the connection). For production use, mercure should not be installed in a VM (the Vagrant setup is only intended for testing and development work).