Single-System Configuration for Clustered Storage Backup
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data management applications struggle to effectively manage backup and restore operations in clustered storage systems due to changes in physical storage configurations, leading to invalid data redundancy policies and loss of traceability for logged events and backed-up data.
Innovation Solution
Implementing a single-system configuration based on logical storage entities within a logical storage environment, which shields physical configurations and allows data redundancy policies to be configured independently of physical changes, using virtual servers, virtual volumes, and virtual backup devices, and utilizing Network Data Management Protocol (NDMP) commands for seamless data management across cluster nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data redundancy policies are configured based on physical storage entities and backup devices, then the DMA can perform backup/restore operations using specific physical configurations, but the policies become invalid when physical configurations change (storage servers moved, renamed, deleted, or tape drives inaccessible)
Solution Approach 1:
The patent introduces a namespace as an intermediary layer between the DMA and physical storage entities. The namespace provides stable logical identifiers (e.g., /mnt/storage/volume1) that remain valid even when physical configurations change. The DMA configures data redundancy policies based on these namespace paths rather than physical entity identifiers, allowing policies to persist across storage server moves, renames, or failures while the underlying physical configuration changes.
2Productivity
If data redundancy policies are updated to reflect new physical configurations, then backup/restore operations can proceed with new storage entities or devices, but previous logged backup/restore events and backed-up data become untraceable
Solution Approach 1:
The patent creates a logical copy or mapping layer through the namespace that preserves the identity of storage entities independent of their physical location. When storage entities are moved or renamed physically, their namespace paths remain unchanged. This allows the DMA to continue tracing logged backup events and backed-up data through the stable namespace identifiers, while simultaneously enabling backup operations to proceed against the new physical configurations through the same namespace interface.
3Reliability
If multiple individual storage servers are integrated into a clustered storage system for load-balance and fail-over capabilities, then storage availability and performance improve, but conventional DMA cannot configure data redundancy policies to take advantage of cluster benefits and may be crippled by special physical configurations
Solution Approach 1:
The patent implements a universal namespace interface that works across individual storage servers and clustered storage systems alike. The namespace provides a consistent logical view whether storage entities reside on single servers or distributed clusters. This allows the DMA to configure data redundancy policies using the same namespace-based approach regardless of whether the underlying infrastructure is a simple storage server or a complex clustered system, enabling the DMA to take advantage of cluster benefits (load-balance, fail-over) without being crippled by the physical complexity.
Data Source
AI summary
A method implements a single system configuration for backing-up and restoring a clustered storage system. The clustered storage system contains a plurality of storage servers. A virtual volume is configured based on one or more of the plurality of storage servers. The virtual volume shields its physical implementation from a client system. A virtual backup device is configured based on a physical backup device connected with one of the plurality of storage servers. The virtual backup device shields its connectivity from the client system. The method backs-up volume data stored in the virtual volume to the virtual backup device, irrespective of which of the plurality of storage servers is utilized by the virtual volume, or which of the plurality of storage servers is connected with the physical backup device.


