Our setup for our PI system has our interface residing on a machine in the same subnet as our DeltaV OPC DA server. Our interface node is running on a physical machine (no VM), and the OPC server is running on a VM App Station maintained by the DeltaV network host. To connect the two servers, we are using the Matrikon Tunneller software. Below are the versions of the different pieces...
PI Server: 2012 x64 (3.4.385.77)
OPC DA Interface: 126.96.36.199
Its not a large system, only one interface and about 10,000 tags. Twice in the past week, the OPC server VM was rebooted. Both times, the OPC interface failed to pick up the OPC points after the reboot. It would not start collecting data until we restarted the interface (for the first incident), and rebooting the entire machine after the second incident.
Should I expect the OPC interface to be more robust that what we are seeing? I had hoped it would be able to just keep working after an OPC source server restart? Is there anything in the interface config that I would possibly look over and/or adjust?
One thing we have talked about to increase its robustness is the possibility of moving the interface onto the same machine as the OPC server, eliminating the tunneller. Does anyone have any thoughts on if this would help? Or just general thoughts on the idea itself?
Please let me know if you would like any additional info about our setup... happy to provide it.