Multiple Network Interface on Communication

Hello. We have a project that includes more than 400 IP based devices.
In Ignition server, we have 4 different NICs.

Can we select different NIC Interface on communication settings in Ignition?
(Because of perfomane issues) or
What do you suggest us for separate network traffic ?

If your NICs are on separate subnets, they will get the appropriate traffic. If you have multiple NICs on the same subnet, you should consider bonding (link aggregate grouping) instead of independent IP addresses.

2 Likes

In addition to what Phil said, we recently went through all of Ignition’s TCP-based device drivers where this would be appropriate and added a “Local Address” setting that can be used to choose a NIC by setting the local bind address to its assigned IP address.

That said, I don’t think you need to worry about it. I have not seen any system with 400+ devices that would generate more traffic than a single NIC can handle. The only time I’ve seen Local Address need to be set is on Windows systems with broken/overlapping routing tables that send traffic out on the wrong NIC.

3 Likes

Something I just fixed in our Ignition installation was the removal of these annoying messages from the logs:

We would get these every minute from all TCP connections, so that cluttered up the logs pretty badly. It was also interesting that the devices were still showing Connected even while logging errors. I believe the problem was that NIC2 on our Ignition server was the one that connected to the rest of our network while NIC1 was unplugged with no IP address settings whatsoever. I checked the routing table using route print and it looked normal (ie. no mention of any other IP besides localhost and the private IP of our Ignition server). The solution was to add the Ignition gateway’s IP address to the Local Address field for each TCP device connection. This removed those pesky error messages.

This could be useful to someone else troubleshooting Error connecting to IP: Connection refused: connect issues.

EDIT: Those pesky log messages are back… I configured the Local Address for all of our TCP devices and those errors in the log went away immediately. However, we had gateway issues and decided to update to the latest version of Ignition and, alas, they are back. This is not a problem functionally but does clog up the logs. I would like to figure out why I get a connection refused: connect error. Any ideas @pturmel @Kevin.Herron?

EDIT 2: I found out that what cleared those error messages (Error connecting to IP. Connection refused: connect) wasn’t adding the Local Address to each of the TCP connections. If you go into the TCP device and just hit the Save changes button without even changing anything, these messages will go away from the logs for some time. Eventually, they reappear. I’m going to keep investigating and will post a solution to this issue if I find one.

It so happens that I never use that driver. I use java sockets instead.

Adding my observations while the situation is fresh in my mind:

  1. I can duplicate this issue by simply restarting Ignition.

  2. When the errors occur the devices (RFID card readers and bar code scanners) can send data to Ignition, but I am unable to write to the devices from Ignition.

  3. I also get inactivity reconnect messages, even thought the devices clearly have activity.

  4. Restarting the TCP device after Ignition has restarted clears the error.

1 Like