Worklist stops working

[Paweł Michałowski]

We are running dcm4chee for ultrasound purposes. The service (dcm4chee-2.18.3) runs on a machine with 4 CPUs, 8GB RAM, CentOS Linux 7.2.1511, PostgreSQL 9.2.14. We have 6 ultrasound machines connected to server.

Every once in a while worklists on ultrasound machines stop working for about 7-10 min. and there is "server is offline" message. At the same time we test server with free tool TestSCU.exe which returns following messages when testing different functions:

• Association Failed! WORKLIST failed to connect to: USGSTORE (10.1.10.27:11112)
• Association Failed! FIND failed to connect to: USGSTORE (10.1.10.27:11112)
• ECHO failed to connect to: USGSTORE (10.1.10.27:11112)

Browsing through dcm4chee-web3 works, ping works, telnet works.

In server log file, after worklist starts working again, we get:

2016-10-18 08:49:50,408 ERROR -> (TCPServer-1-1355) [org.dcm4cheri.net.ActiveAssociationImpl] Read timed out

java.net.SocketTimeoutException: Read timed out

We have similar setup for our X-Ray machines on different server and it works ok.

Please point me in the right direction, what to look for or what to test to solve this problem.