Using Edge as a remote OPC UA connection (Redundancy issue)

Ignition
1

We are using a redundancy setup.
The Ignition Main Gateway is configured to connect to the built-in OPC server of the remote edge gateway, allowing us to access local devices on the edge through this new OPC server.

We have enabled and configured both the Failover functionality in the OPC Client and Redundancy in the Server, as we have a redundant edge server setup. However, we are encountering the following issue:

  • When we manually force a failover from the master edge to the standby server, the OPC connection remains linked to the master edge server, even though it is in a Cold (inactive) state.
  • If the main edge server machine is completely shut down, the failover works as expected.

Is this behavior intended? If we want to provide external access to the data, we need the OPC connection to always switch to the active server in the redundant setup.

2

Yes, however much less than ideal it is.

The janky "failover" concept on OPC Connections isn't connected in any way to either Ignition redundancy or to OPC UA redundancy concepts. It simply looks at whether the connection is up or not, and then switches to the failover configuration when it hasn't been up for long enough.

Eventually we'll implement support for OPC UA Non-transparent Redundancy in a new connection type, which means even if the master stays online the client will be aware of the redundancy role and switch accordingly, including switching back when the original master resumes control, which is something the "failover" thing also doesn't handle.

1 Like
3

Thank you, Kevin, for the quick and clear response.

I have to say, it’s disappointing to hear that the current "failover" implementation isn’t integrated with Ignition redundancy or OPC UA redundancy standards. This limitation makes it challenging to rely on the setup for seamless external access to data, which is a critical requirement for us.

It’s good to know that support for OPC UA Non-transparent Redundancy is planned for the future, as that would address these shortcomings. In the meantime, it seems we’ll need to explore alternative solutions to maintain reliability.

Thanks again for your swift reply and for providing clarity on this issue.