Master/Backup Tag Provider Architecture - Single Point of Failure

Ignition
1

I'm running into a reliability issue with our current tag provider architecture and looking for suggestions. Here's our setup:

Current Architecture:

  • Remote Site provides tags via remote tag provider
  • Master node receives these tags and acts as an intermediary
  • Backup Node receives tags from Master
  • We have Frontend Gateways getting tags as well through remote tag providers
    Tag Flow: Remote Site -> Master/Backup -> Frontend

Problem:
When the Master node goes down, we lose all tag readings. The core issue is that the tag provider on our Master and Backup is itself a remote tag provider, creating a single point of failure. We cannot connect our Frontend Gateways directly to the Remote Sites.

Any suggestions would be greatly appreciated!

2

Is your "frontend" just clients of the master? Or is it another gateway? I'm thinking you've got gateways, but my question would be why?

In any case, with this sort of setup any of your remote tag providers are a point of failure. This is biggest downfall of remote tag providers and why I chose a hub and spoke architecture with redundant tag connections rather than remote providers.

3

The frontends are other gateways (frontend-vision and frontend-perspective). The reason for this setup was to better handle resource usage and security.

Your point about hub and spoke with redundant connections is interesting. Could you elaborate on how you implemented this? Given that we can't modify the backup node tag providers, would there be a way to transition to a hub/spoke model while keeping the backup node as is?

4

Well you can run multiple projects on a single gateway and have vision and perspective, but I digress.

If your master goes down, do you not have the backup there for it? So, are you working on a solution for if the master and backup are down?

5

To clarify our setup:

  • We have Master and Backup gateways configured as a redundant pair (using standard Ignition redundancy)
  • A remote tag provider is configured on the Master gateway, which automatically replicates to the Backup
  • Frontend Vision and Perspective are separate gateways connecting to this redundant pair
  • When Master goes down, the Frontends immediately switch their 'connection' to the Backup but can't get tag data from the Backup (even though they should)
  • I think if the tag providers on Master/Backup were standard tag providers, this would have worked. But, because we are using remote tag providers, this does not work
6

I'm going to guess that there is something about your Master/Backup network environment that isn't matching, or perhaps there's a GAN connection that was never approved.

1 Like
7

It works now. I made an outgoing connection from the Remote Site to both Master and Backup. Then, I made an outgoing connection from the Frontends to both Master and Backup.
Visualized:

image
image992×578 26.4 KB

1 Like