Redundancy failover alternate address configuration

Ignition
1

I have a gateway network setup where a master gateway and redundant gateway are connected through the gateway network. They are configured correctly as “master” and “backup” and both show “peer connected” in the gateway settings webpage. There are scenarios where a client can only connect to these two gateways through an Nginx proxy. Here’s an example where master and backup gateways are running on different VMs and the Nginx proxy that handles both is running on a separate VM:

master GW: 192.168.12.10:8088, backup GW: 192.168.12.11:8088

master GW proxy: 10.20.30.10:1010, backup GW: 10.20.30.10:1011

My issue is when a vision client is connected to the master at 10.20.30.10:1010, and a failover is triggered, and the 192.168.12.11:8088 address is not reachable, the gateway (I suppose obviously) doesn’t know to redirect the client to the proxied address (10.20.30.10:1011). This is not an issue when the client is connected to the non-proxied address. Is there a gateway setting I’m missing that would allow for this behavior? Thanks in advance!

2

No, but this can be solved by using DNS names instead of IP addresses, and ensuring that the DNS infrastructure performs localized queries (returns the IPs appropriate for the requestor's subnet). Set the DNS names as the gateway's public addresses.

Using DNS names is one of my top recommendations to clients. (Especially when you want to enable HTTPS.)

1 Like