High Availability Demo on an OPA System

Show notes

This episode is a live demo of high availability in an O-PAS-based automation system, with Jarie hosting alongside Andrew and David. They walk through how the setup keeps a process running through device changes and failures, and why that matters for real industrial environments where downtime is costly.

They discuss how O-PAS extends redundancy concepts, how the demo system is wired, and what happens when controllers are unplugged and reconnected. The conversation also shows why software-driven high availability can support patching, migration, and mixed hardware without stopping the process.

Watch on YouTube. Listen on Spotify or Apple Podcasts.

Key topics

• Andrew explains the difference between redundancy and high availability, and why HA is about the uptime target rather than just having duplicate devices.
• David explains why process industries need HA when they cannot tolerate downtime from failed processors, control logic, or other equipment.
• The team describes O-PAS changes to the spec that add more detail around OPC UA redundancy and HA, including the distinction between informative and normative requirements.
• Jarie walks through the live demo setup, including the COPA 500 system, the simulated process with a ping pong ball setpoint at 50, and the engineering workstation.
• The demo shows a mixed-vendor environment with CODESYS control logic, Ignition for HMI, Phoenix I/O, Phoenix power supply, Cisco and Netgear switching, and ASRock-based DCN units.
• They demonstrate how opening and closing a valve changes airflow, which causes the ball level to move and then recover through PID control.
• The team simulates DCN failover by unplugging controllers and watching the HA status switch between leader and follower.
• David and Andrew explain that the system can scale beyond two controllers, including three or more DCNs for maintenance, patching, and migration scenarios.
• They note that virtual DCNs are possible and that HA reduces tight hardware/software coupling by allowing different hardware models to work together.
• Andrew explains that leadership can be forced through software if needed, including within the control runtime or via Fusion.
• David highlights that failover happens on the next scan cycle without a resynchronization scan, so the promoted follower takes over immediately.
• The episode closes by reinforcing that O-PAS-based interoperability and high availability can reduce downtime and improve maintainability.

Timestamps

• 00:00 — Live demo introduction and O-PAS high availability overview
• 00:59 — Andrew defines high availability vs redundancy
• 03:43 — David explains why HA matters in process industries
• 06:06 — Why HA helps with patching, replacement, and planned downtime
• 07:13 — Walking through the demo hardware and process simulator
• 09:04 — Software interoperability with CODESYS and Ignition
• 10:02 — Simulating valve changes and controller failover
• 12:36 — First successful process disturbance and recovery
• 13:24 — DCN1 unplugged and leader/follower role changes
• 14:07 — Reconnecting controllers and recovering the system
• 15:37 — Scaling HA beyond two controllers and supporting migrations
• 16:12 — Virtual DCNs and reduced hardware coupling
• 17:59 — Forcing leader selection through software
• 18:21 — Priority-based leadership and failover behavior
• 20:50 — Next-scan-cycle takeover with no resynchronization delay
• 21:34 — Wrap-up and closing remarks