The 3 keep-alive config properties, thus keep-alive frequency, keep-alive timeout and keep-alive failures allowed are missing from the documentation. Instead, there is a “Secure Channel Re-authentication Enabled” config property.
1 Like
I noticed this as well. There’s also not such a good explanation on this page as to the differences between backup and failover endpoints. If you have redundant Ignition systems and a redundant third party OPC server should you configure both?
Backup endpoint configuration determines what OPC server the backup Ignition gateway will connect to - because you (probably) want it to be different on the master and backup, especially if they're using local servers.
Failover is for the OPC connection that will be used (by either the master or backup) if their primary connection fails.
If you don't specify a backup endpoint, then the master and backup Ignition system will both prefer the primary connection, unless it fails, in which case they will switch to a failover (if it's populated).
You have to control all those settings from the master gateway because the backup gets a direct copy of the settings you populated on the master.
Side note: the extra configuration for the backup only exists because once upon a time if you had configured an actual IP address for the server to bind to then the "loopback" OPC UA connection also had to use that IP address instead of being able to use "localhost" in the hostname. I'm pretty sure that's no longer the case in 7.9.x, and it's definitely not the case in 8.0
OK, so this is relevant only if you want the backup Ignition gateway to use a different OPC endpoint than the primary would? Would the failover endpoint not still work for this though in most cases?
Example:
Primary server running Ignition and Kepware VMs
Backup server running Ignition and Kepware VMs
-
Primary server goes down so Ignition switches to backup. The OPC server also went down and is unreachable so Ignition switches to the failover endpoint.
-
Primary Ignition VM crashes but the server is fine. Ignition switches to the backup server but continues to communicate with the primary's OPC server.
It sounds like this would be useful mainly to get Ignition to "prefer" connection to another OPC server while it's running in backup mode. Maybe the backup server is located in an outside datacenter and lag time would be significantly increased if the backup Ignition continued using the primary OPC server.
It could also be important if you had specified an internal network location "localhost" or something in the settings but where this would not signify the same location from the perspective of the backup Ignition gateway. Correct?
Yes, correct on both counts. We have an active ticket to look at supporting more complex failover/redundancy strategies for OPC, but it still needs some investigation/refinement.