Troubleshooting PI System AF Server Analysis Failures

Troubleshooting PI System AF Server Analysis Failures

In the world of data management and analytics, OSIsoft's PI System is a powerful tool for capturing and analyzing operational data in real-time. However, like any sophisticated system, it can encounter issues that require troubleshooting. A common problem is when AF Server analyses suddenly stop working, a scenario that can puzzle even the most seasoned engineers. This blog post will delve into potential causes and solutions for such an occurrence, inspired by a real situation shared by a process engineer.

Identifying the Problem

In this instance, a process engineer added new analyses to existing elements within the Asset Framework (AF). Following this change, several older analyses associated with those elements stopped functioning correctly, despite all indicators showing they were running. Notably, the element that wasn't modified continued working seamlessly. This peculiar behavior raised several questions:

• Why did the problem affect only those elements with new analyses added?
• Could the AF server be overburdened by excessive computation?
• How to verify if the server's hardware is sufficient for the given computational load?

Understanding the System Behavior

When new analyses are introduced and checked into an element, all analyses for that element are reinitialized. This process is crucial for integrating the new analytic processes. However, reinitialization can consume system resources and potentially lead to temporary halts in computation, especially if the system attempts a backfill to cover the transition period.

Another consideration is that analyses may lag if the server cannot process inputs at the required pace. Lagging analyses might show as active within the system’s interface, yet fail to reflect real-time results.

Solutions and Best Practices

1. System Information and Computation Load:

   • Monitor AF Server Performance: Utilize the PI System Management Tools (PI SMT) to check resource utilization. Monitoring CPU, memory, and network usage can reveal if the server is operating at capacity.
   • Check Analysis Lagging: Use Performance Counters in Windows to assess whether there are delays in processing analysis data.

2. Hardware Adequacy:

   • Capacity Planning: Evaluate if your hardware aligns with the number of analyses and volume of data being processed. OSIsoft provides guidelines on hardware specifications, suited for data throughput and storage requirements.
   • Tech Support Assistance: OSIsoft offers technical support which can provide specific insights into potential system bottlenecks and recommended hardware configurations. Engaging their expertise is beneficial for complex troubleshooting.

3. Management of Reinitialization and Backfills:

   • Stagger Changes: Consider introducing changes in smaller batches to minimize system impact, allowing analyses time to stabilize post-reinitialization.
   • Optimize Analysis Configuration: Simplify or reduce the frequency of analyses where possible. Event-triggered analyses can be particularly resource-intensive if not optimally configured.

Conclusion

Mystifying analysis failures in the PI System's AF server can often be attributed to computational strain from new configurations. To mitigate these issues, implement performance monitoring, ensure adequate hardware, and consider phased updates. Regular system audits and optimization can preclude many common pitfalls, maintaining robust and real-time data analytics for your operations. By adopting these strategies, you can ensure that your PI System continues to serve as a reliable foundation for your operational data needs.