Migrating PI Server: Strategic Minimization of Downtime

When migrating a PI Server, especially within a Rockwell FT Historian environment, it's crucial to approach the process methodically to minimize downtime and ensure data integrity. Here, we will discuss a scenario involving an upgrade from PI Server V3.4.390.18 running on Windows Server 2008 to PI Server V3.4.45.604 on Windows Server 2019 High Availability (HA) setup.

Challenges and Considerations

The migration discussed in this blog poses several unique challenges, such as:

1. Version and Compatibility Restrictions: The existing Rockwell FT Historian (PI Server) version is compatible only up to Windows Server 2008 R2 SP1, which requires a strategic approach since the new target environment is Windows Server 2019.

2. High Availability Setup: Transitioning from a standalone server to an HA configuration necessitates additional considerations to ensure that the configuration settings and data synchronization between nodes are handled correctly.

3. Minimal Downtime Concerns: The critical objective in most industrial environments is to minimize downtime, as any delay can lead to significant operational and financial repercussions. Hence, Vikas’ plan to restore backups now and execute archive restores at the time of cutover is particularly clever.

Step-by-Step Migration Plan

• Backup and Initial Restoration:

  • Begin by taking a comprehensive backup of the existing PI Server. This backup will serve as the foundational data set to initialize your new HA servers.
  • Deploy this backup on your new Windows Server 2019 environment, while ensuring that you do not start using the newly set up servers until the cutover to avoid any data sync issues or loss of data input.

• Upgrading PI Server Version:

  • Due to the incompatibility of the older PI Server with Windows Server 2019, it's necessary to perform the upgrade after initial deployment. This can be done by either installing the older version on a compatible environment for data transfer or upgrading the application post-backup restore.

• HA Collective Setup:

  • Once the backup is successfully restored and the PI Server version is upgraded, proceed to set up the High Availability collective as per Rockwell’s specifications for FT Historian.

• Cutover and Archive Restoration:

  • On the designated cutover date, take a final archive backup from the current production server. Restore this latest archive onto the new HA setup.
  • Begin redirecting live data streams and client applications to point to the new HA environment, thus ensuring continuous data collection with minimal downtime.

Conclusion

Migration of critical infrastructure such as a PI Server requires careful planning and execution. By considering OS compatibility, data synchronization needs, and employing strategies to minimize downtime, this transition can be achieved smoothly. This migration strategy not only aligns with industry best practices but also highlights the flexibility needed to adapt to technological advancements while maintaining operational integrity.