VSS Writer Metadata Modification for Cluster Shared Volume Restore
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies lack efficient methods for restoring backups of cluster-shared volumes, particularly in high availability cluster environments, where application-specific writers do not report data on remote nodes, leading to incomplete or unsupported restore operations.
Innovation Solution
Modifying the writer metadata document (WMD) and backup component document (BCD) to allow restoration operations to involve application-specific writers without invoking the CSV writer, enabling restoration without requiring CSV awareness, thus supporting legacy restore operations for applications like SQL and Hyper-V deployed on CSV from a standard LUN.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If application-specific writers do not report data on remote nodes, then the backup process can be simplified, but restore operations become incomplete or unsupported
Solution Approach 1:
The patent introduces a CSV writer as an intermediary component that bridges application-specific writers and remote nodes. The CSV writer reports data from remote nodes to the backup system, enabling complete restore operations while maintaining the simplicity of application-specific writers. This mediator resolves the contradiction by adding a specialized component only where needed for restore operations.
Solution Approach 2:
The patent segments the backup and restore functions into separate phases with different requirements. During backup, application-specific writers operate independently for simplicity. During restore, the CSV writer is invoked to provide remote node data, ensuring completeness. This temporal and functional segmentation allows each component to optimize for its specific phase without compromising the other.
2Reliability
If CSV writer is invoked for restore operations, then restore completeness is improved, but system complexity increases
Solution Approach 1:
The CSV writer is designed with multi-functionality, serving both as a coordinator for backup operations and as a provider of remote node data for restore operations. By making the CSV writer universal, the patent avoids adding separate components for restore functionality, thereby limiting the increase in system complexity while maintaining restore completeness.
3Adaptability or versatility
If legacy restore operations are supported from standard LUN, then compatibility is improved, but support for cluster-shared volumes is limited
Solution Approach 1:
The patent creates a virtual copy of the CSV writer's reporting functionality that can be invoked during restore operations. This virtual copy enables legacy restore operations to work with CSV data without requiring actual CSV awareness in the restore path, thus maintaining compatibility while supporting cluster-shared volumes.
Data Source
AI summary
A first cluster node receives a request for a backup of files stored in a CSV shared with a second cluster node. A first VSS writer determines a first set of files associated with the first cluster node. A second VSS writer determines a second set of files associated with the second cluster node. A WMD file associated with the backup is generated to include a first component describing the first set of files and a second component describing the second set of files. A BCD file associated with the backup is modified to store information indicating the first VSS writer to restore the first and second sets of files, such that the first and second sets of files can be restored using the first VSS writer without having to use the second VSS writer. The backup is then performed to a backup storage device.