800xA 6.0.3 Harmony Connect fails writes on redundant server intermittently
We have had a lingering issue with our v6.0.3 Harmony Connect servers. Every few months the server set as 'Inactive' is unable to perform PUT operations (SP, CO Changes). The GET operations seem to be working fine as data is still changing. (You can see the SP change from a node connected to the other server). The 'Active' server is still able to perform these operations. On much rarer occasions, it has affected both servers. The current fix is to take the affected server's 'Opc_DA_Connector' out of service and put it back in service, or to reboot the affected server.
This type of message is visible in the diagnostic file each time.
<DiagObj ID="Server still busy processing previous operation, put operation cancelled">4/9/2019 12:48:54 PM (.824)</DiagObj>
Has anyone seen this before? What steps can I take to view deeper diagnostic information?
This question has not yet been answered.