Hey guys, probably something silly I am doing wrong but just testing an OPC client connecting to Ignition’s OPC server. Everything works well on the main gateway but I am having a little trouble when I fail over, the server is still connected fine and appears to be able to read the test tag (in this case just a memory tag).
However, on further on inspection the read although a good quality does not update. What is adding to my confusion i guess is that i can actually write from the client successfully and although it doesn’t update, if I fail back to the main gateway it then displays the modified value. Where could I be going wrong on failover if the reading isn’t updating but can still write? Also I am using the latest version of Ignition 8.0.15
We are developing out an app in 8.0.13 and have had to write our own wrapper function around the system.opc.write. Eventhough the result comes back as isGood the value never makes it down to the plc and we have had to implement retry by doing a “read after write” in our wrapper to then retry. We are convinced there is some kind of bug buried in the OPCUA module as of version 8.0.13.
Frank
@anon55648468 I’m not sure I understand the scenario - are you leaving your 3rd party client connected to the primary server when you force a failover? Which server is active and which is your client connected to when you do the writes?
@franklepkowski your issue sounds entirely different than this. This issue comes up from time to time and it has never been anything other than an issue on the PLC side. Whether it’s logic in the PLC resetting the value or a PLC simply not updating the value even though the write returned successfully. Maybe you can capture this happening on Wireshark if it’s frequent enough and we can verify.
Hi Kevin, that is correct. I have left the client connected to the primary and I am able to write through this connection after failover but not recieve any updated reads.
@anon55648468 I’m not sure what the expected behavior here should be, I’ll have to get back to you.
I think it’s not really surprising that the reads are stale, but the writes should probably fail. Maybe the reads too.
Cheers Kevin, I would have thought the read/write would return a bad quality (so to indicate to the client that not all is well and give it a chance to repoint to the backup server) and opc ua server redundancy feature would only kick if the actual ua server failed, which it will not do in this test because the machine is still operating normally
Hi,
Is there any update on this issue, I have the same situation with my client, the opc ua client will switch to the backup gateway opc ua server and remain connected even if the master becomes active (the backup opc ua server remain active with no indication of failover).