Automatic Snapshotting for VM Instance Store Volume Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud-based service providers face challenges in ensuring data persistence and recovery during virtual machine instance migrations, particularly when instance store volumes are involved, as data loss can occur if the migration to a new host is unsuccessful, leading to unrecoverable customer data.
Innovation Solution
Implementing a snapshot component within the service provider network that automatically obtains point-in-time snapshots of instance store volumes before migration, storing these snapshots in a network-based storage service, allowing for data recovery and potential retry of migration without user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If live migration of VM instances is supported for instances with instance store volumes, then service continuity and availability are improved, but data loss risk increases because data in instance store volumes may be lost if migration fails
Solution Approach 1:
The system performs a preliminary snapshot of the instance store volume before initiating migration. This snapshot is stored in persistent storage as a safety mechanism. If migration fails, the data can be recovered from the snapshot, thus preventing data loss while enabling migration of instances with local storage volumes.
Solution Approach 2:
The system creates a protective snapshot copy of the instance store volume data before migration occurs. This snapshot acts as a cushion or safety net that prevents complete data loss if the migration process fails, allowing the system to withstand migration failures without losing customer data.
2Reliability
If snapshots of instance store volumes are automatically obtained before migration, then data recovery capability is improved, but system complexity increases due to additional snapshot management operations
Solution Approach 1:
The system automatically detects when a VM instance has an instance store volume and autonomously creates snapshots before migration without requiring customer intervention. The snapshot management is self-service, handling creation, storage, and potential recovery operations automatically, which reduces the perceived complexity for users while maintaining robust data protection.
Solution Approach 2:
The snapshot acts as an intermediary mechanism between the instance store volume and the migration process. It provides a decoupled safety layer that manages data protection separately from the migration operation, allowing the migration system to proceed with confidence while data recovery capability is maintained through the intermediary snapshot storage.
Data Source
AI summary
Snapshots of instance store volumes may be automatically obtained prior to an upcoming migration event for an associated VM instance(s). This automated snapshotting allows customers to restore their data from the snapshot of their instance store volumes in the event of a failure during the migration event that results in an inability to start (or launch) the customer's VM instance on a target host computer. In an example process, a computing device(s) may determine that a migration event is scheduled to be performed for a VM instance, determine that the VM instance is associated with one or more instance store volumes, obtain a snapshot of the one or more instance store volumes, and store the snapshot in storage of a network-based storage service prior to performing the migration event. In some embodiments, the snapshot may be stored in association with a service provider account of the network-based storage service.