Faulty Connection: NJS-003 caused by DPI-1010 after ORA-03113

[nauben]

Describe the bug
The Node-RED system is placed behind a firewall that closes open connections from Node-RED to the Oracle database after a certain amount of time when no requests are made. As a result the error ORA-03113 occurs. Thereafter, all attempted queries result in the error NJS-003 which in turn is caused by DPI-1010.

Oracle query error: NJS-003: invalid or closed connection
DPI-1010: not connected, errorCode:NJS-003: 

Expected behavior
Error should be catched, the current connection should be closed and a new connection should be established. Ideally, this already happens at the error ORA-03113; but at the latest on the consecutive NJS-003 / DPI-1010

Desktop (please complete the following information):

• OS: Docker Image (Debian) on Kubernetes Cluster
• Node-RED Version 3.0.2
• Node.js Version 16

Additional context

• While the sequence of errors is the same every time, it is just speculation that the firewall caused this or the Oracle database itself.
• The automatic reconnection feature unfortunately does not help.
• When the node is copied and the original deleted while the new node replaces it, then the connection to the database works again.
• When the container is attempted to be stopped by executing ```kill -s SIGINT 1```` from internally, it can take a very long time (up to 45 min) to finally be stopped.
• Adding ?expire_time=1 to the TNS name helps to delay this issue. After some time however (maybe another error) the same behaviour with NJS-003 occurs.