The backup process was incorrectly configured and performed an unintended cluster node downgrade/removal, violating the requirement for an odd-numbered cluster size.
Impact:
- Service instability and intermittent outages
- Elevated risk of data inconsistency (no permanent data loss confirmed)
- Degraded system performance during incident window
Resolution:
- Affected node was restored and cluster returned to an odd number configuration
- Cluster quorum and leader election normalised
- System integrity verified
Address Root Cause Actions:
- Update backup processes to ensure disconnecting node approach is no longer utilised on backup nodes
- Enforce configuration policies requiring odd-numbered node counts
The above actions will ensure the root cause will be addressed.