Virtual Machine Restoration from Datastore Copy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for restoring virtual machines from a datastore in a disaster scenario often require recovering the entire datastore, which is time-consuming and inefficient, especially when only a single virtual machine is affected by corruption or failure, impacting other virtual machines and prolonging recovery time objectives (RTO).
Innovation Solution
A method that allows for the selection and restoration of a specific point-in-time copy of a virtual machine, involving the retrieval of catalog information, mounting a datastore, modifying the virtual machine configuration, registering the modified path, determining the snapshot ID, and reverting to a snapshot of the virtual machine, enabling rapid recovery of individual virtual machines within minutes regardless of datastore size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the entire datastore is recovered to restore a corrupted virtual machine, then the virtual machine can be restored, but the recovery time is excessively long and other virtual machines are unnecessarily impacted
Solution Approach 1:
The patent segments the restoration process by separating the virtual machine configuration files from the bulk datastore. It extracts only the specific virtual machine's configuration files (vmx, vmdk, etc.) from the mounted datastore, rather than recovering the entire datastore. This segmentation enables selective restoration of only the affected virtual machine components, dramatically reducing recovery time from hours to minutes while avoiding unnecessary impact on other virtual machines.
2Reliability
If the entire datastore is recovered to restore a corrupted virtual machine, then the virtual machine can be restored, but other virtual machines are affected and recovery efficiency decreases
Solution Approach 1:
The patent applies the extraction principle by removing only the necessary virtual machine configuration files from the mounted datastore, rather than recovering the entire datastore. The system identifies and extracts specific files (vmx, vmdk, nvram, log files) associated with the corrupted virtual machine, leaving other virtual machine data intact. This selective extraction approach restores the affected virtual machine without disrupting other virtual machines, thereby maintaining high disaster recovery efficiency.
3Loss of time
If selective restoration of individual virtual machines is implemented, then recovery time is reduced, but the complexity of the restoration process increases
Solution Approach 1:
The patent introduces an intermediary approach by mounting the datastore as a virtual filesystem and using configuration file path modification as a mediator between the restored virtual machine and the original datastore. The system modifies the vmx file's diskPath references to point to the mounted datastore location, creating a seamless connection without requiring complex data migration or system reconfiguration. This intermediary mechanism simplifies the overall restoration process while enabling rapid selective recovery.
Data Source
AI summary
In one aspect, a method to restore a virtual machine comprising receiving a selection of a point-in-time copy to restore, retrieving catalog information associated with the point-in-time copy of the virtual machine to restore, mounting a datastore on a host, modifying a virtual machine configuration file path, registering modified virtual machine configuration file path with the host, determining a snapshot ID of a snapshot of the virtual machine and reverting to the snapshot of virtual machine.