OPC Server(5500) Connection error to DA subscribed controller
In a 800xA RevB FP3 (64-bit), I'm getting the system alarm "OPC Server(5500) Connection error to DA subscribed controller" for a redundant PM866. It happens both randomly and when downloading to the controller from CBM in the step "Analyzing changes". When CBM is closed, the controller status goes back to normal.
Controller has 1 CI854 redundant module with 4 CI840 redundant slaves. Cyclic charge is under 20%.
System has DC, AS1, AS2, CS1 (node 21), CS2 (node 22) and IM nodes.
When it happens randomly it lasts several minutes and even hours until it returns to normal, it's not possibble to go online, download or open Remote System Dialog. But operator workplace doesn't lose data access or alarms. Also in RNRP controller shows (up, up). We realize that closing CBM fix the issue until next download.
When downloading I get ther error on the step "Analyzing changes". OPC Servers shows X on Data Access tab and operator workpalce lose data after "you are now online" step so I have to close CBM to get data back.
In controller log I get the line:
0x138b4e0 (Main): arpresolve: lookup failed (resource shortage or network configuration error -- check netmask), among others. Log is attached.
Each Control Network is distribuited over 3 switches (1 managed, 2 unmanaged). In managed switch I have fixed 10 Mbps HD for controllers and 100 Mbps FD for connectivity servers and port that links with unmaged switches. In both CS control network NICs are fixed at 100 MBbps FD.
I have tried downloads with only 1 CPU active, 1 CS active, only primary network active, only secondary network active, CBM from different nodes but still have the same result.
I would really appreciatte if you can help me.
Voted best answer
"arpresolve" is presumably the process that resolves dns names and/or IP adresses to MAC adresses. And errors here would most likely be the cause of lost communication. As the message suggests, start by looking for network configuration errors on your controller, engineering station and connectivity servers.
Its also worth looking for errors in your DNS server.
Around the time of "Analyzing Changes" control builder will try to send updated configuration files from the local CBM directory to the remote OPC server config files directory. If something is wrong in these files or if the files dont get sent properly, then the OPC connection between the CS and the controller will fail. So my first guess is that something is going wromg here because of bad address settings, but the system is somehow robust enough that it doesnt shut everything down. Possibly its your redundant CS thats keeping everything running ?