OPC Surrogate on a network disconnect.
We setup a new application server for PI Interface and customer was trying to perform a redundancy test and on a network disconnect the OPC Surrogate freeze until the network connection returns back to normal state.
This makes the PI interface OPC to think like the frozen connection was good and didnt failover.
My question was why OPC Surrogate connection not timed out on a network disconnect.
Voted best answer
Stop using the PI-OPC interface and use the OPC-HDA HISTORY Interface instead.OSI have a standard history interface. I belive the licence costs are the same so your customer should be able to exchange the interfaces.
The history database can recover data after a failure and is much more reliable than the OPC-DA interface. It also doesnt add extra load to the controllers.
The issue you had is with the PI interface. PI thinks its still connected and doesnt failover. You will need to discuss this with OSI. In my experience, the failover on the OPC-HDA interface seems to be much more reliable at detecting service failures.
The PI history interface can be configured using the Log Manager tool in Bulk Data Manager to extract the 800xA history log names.
OK. So the OPC Surrogate Server becomes totally unresponsive and the local workplace is also affected ?
Check RNRP for configuration errors and ensure that the network cable failure is detected correctly.If the Application server is also connected on the PCS network you need to configure RNRP settings for that too. The PCS network is a local network. Make sure RNRP isn't trying to fail over onto the PCS network.
A single network cable failure should be handled by RNRP and the OPC Surrogate server should continue to function. If both Plant network cables fail and the Application Server losses connection to the 800xA Aspect system entirely then the OPC Surrogate server should mark it's tag data quality as BAD.