VM Snapshot Backup Consistency via In-Guest Agent Coordination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current snapshot backup systems in virtualized environments face challenges such as application-awareness protection, time windows for snapshot generation, and support for heterogeneous disks, leading to potential data loss or corruption due to failed backups and incomplete snapshots.
Innovation Solution
A data protection backup system with an in-guest data protection agent and a host-level snapshot data mover that coordinates with the copy service and virtual machine management center to extend snapshot generation time, manage disk types, and ensure application consistency, using techniques like light-weight delta filters and secure communication protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the snapshot generation time is extended to complete application-consistent backups, then the reliability of backups is improved, but the productivity of the backup system deteriorates due to exceeding prescribed time windows
Solution Approach 1:
The system performs preliminary actions by notifying the application to flush memory contents to disk before the snapshot is actually taken. This ensures data consistency is prepared in advance, allowing the snapshot itself to be captured quickly within the prescribed time window without compromising reliability
Solution Approach 2:
The backup process is segmented into distinct phases: a preparation phase where the application is notified and memory is flushed to disk, and a snapshot phase where the actual snapshot is taken quickly. This segmentation allows the time-consuming consistency preparation to occur before the critical snapshot window, improving both reliability and speed
2Reliability
If the application is frozen for snapshot generation, then the application consistency is improved, but the productivity of the system deteriorates due to processing cycle diversion
Solution Approach 1:
The application is notified in advance to flush its memory buffers to disk before the snapshot is taken. This preliminary action ensures that all data is safely written to persistent storage, guaranteeing consistency without requiring the application to remain frozen during the actual snapshot capture
Solution Approach 2:
The system dynamically adjusts the freezing duration to be minimal—only long enough to capture the snapshot after the flush is complete. The application is frozen briefly during snapshot capture but remains operational during the flush phase, optimizing the balance between consistency and productivity
3Productivity
If copy service time windows are shortened to improve backup speed, then the productivity is improved, but the reliability deteriorates due to timeout before snapshot completion
Solution Approach 1:
The system performs the data flush operation before initiating the snapshot, so that when the snapshot window begins, the application is already in a consistent state ready for rapid capture. This preliminary preparation enables snapshots to complete within shorter time windows without timeout errors
Solution Approach 2:
The backup operation is divided into a pre-snapshot preparation phase (flushing data) and a snapshot capture phase. This segmentation allows the snapshot phase to be completed quickly within tight time windows, improving both speed and completion rate
Data Source
AI summary
A data protection agent is installed as a guest within a virtual machine (VM) hosted by a host and a snapshot data mover is installed external to the VM on the host. Upon receiving a command to conduct a backup, a first request is issued to a copy service of the VM indicating that an application in the VM is to be backed up. A second request for a snapshot is issued to the snapshot data mover. The snapshot data mover passes the snapshot request to a virtual machine manager. Upon the snapshot data mover receiving an identifier for the snapshot, the snapshot data mover passes the snapshot identifier to the agent thereby allowing the agent to locate the snapshot and conduct the backup.