Deployment and Rollback Enhancements

We are requesting the IBM Sterling Integrator (B2Bi) development team to implement an enhanced rollback mechanism for non-outage upgrades in environments consisting of multiple clusters.

The current rollback strategy relies on restoring a backup snapshot of the Sterling nodes and the database, which leads to data loss for any activity processed after the backup. This is a critical limitation for enterprises that require uninterrupted data processing and cannot afford any data loss during or after the upgrade process.

Current Limitations:  

1. The current rollback process requires restoring a snapshot of the system, which includes both the Sterling nodes and the associated database.  

2. Any data written to the database during the upgrade is lost upon rollback, leading to unacceptable risks for critical business operations.  

3. There is no seamless or incremental rollback strategy that allows returning to the previous version without impacting data integrity or ongoing business processes.  

We request IBM Labs to develop a robust and reliable rollback mechanism for non-outage upgrades that ensures:  

1.  Data Integrity: 
The rollback process must maintain the integrity of all data processed during the upgrade window, ensuring no data is lost.  

2. Transaction Continuity: 
The rollback mechanism should maintain transactional consistency and allow for a seamless transition to the previous version without impacting ongoing processes.  

3. Incremental Rollback: 
Introduce functionality to roll back only the software changes while preserving any data changes made during the upgrade process. This could include version-aware rollback capabilities or a mechanism to reverse only application-level changes.  

4. Cluster-Wide Synchronisation: 
The rollback process should account for multi-cluster environments where different nodes may be at different stages of the upgrade, ensuring consistent rollback across the clusters.  

5. Upgrade Documentation and Guidance: 
Clearly define in release note documentation whether an upgrade can be performed in a non-impacting, non-outage fashion. Additionally, we request that IBM prioritises and designs upgrades with non-outage scenarios as the default approach, minimising the need for outage-based upgrades whenever possible.

The lack of a reliable rollback mechanism increases the risk of data loss, operational downtime, and disruption of critical business processes during upgrades. This limitation forces our teams to take extreme precautionary measures, such as extended maintenance windows or restricting upgrades to non-critical periods, both of which are impractical in a 24/7 business environment with >99% availability commitments.

A robust rollback mechanism would significantly improve business continuity, reduce risks, and increase the confidence of enterprises in upgrading IBM Sterling Integrator (B2Bi).