Asynchronous Replication Point-in-Time Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data protection systems face challenges in maintaining a consistent point-in-time snapshot of virtual machines across sites, especially during data replication, which is crucial for disaster recovery and data integrity.
Innovation Solution
The method involves pausing write I/Os for a second virtual machine, generating a snapshot and bookmark, resuming I/Os, rolling the journal to a specific point-in-time, and creating a snapshot of the virtual volume at the first site using the journal, ensuring data consistency and recoverability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If asynchronous replication is used to replicate virtual machine data between sites, then data protection and disaster recovery capability are improved, but maintaining a consistent point-in-time snapshot across sites becomes difficult
Solution Approach 1:
The system performs preliminary actions by pausing write I/O operations for the second virtual machine before generating snapshots at the first site. This ensures that no new data changes occur during the snapshot process, guaranteeing that the source and destination snapshots capture the same point-in-time state. After the snapshot is complete, write I/Os are resumed. This preliminary pause action resolves the contradiction by ensuring temporal consistency without sacrificing asynchronous replication benefits.
2Measurement precision
If write I/O operations are paused for the second virtual machine to ensure consistency, then point-in-time snapshot accuracy is improved, but system productivity and data flow are reduced
Solution Approach 1:
The system applies periodic action by pausing write I/O operations only during the brief moment when snapshots are being generated at the first site, rather than continuously pausing. The pause is temporary and synchronized with the snapshot operation duration. After the snapshot is complete, write I/Os are immediately resumed. This periodic, temporary pause minimizes the impact on productivity while ensuring snapshot consistency, resolving the contradiction between accuracy and efficiency.
3Reliability
If a bookmark is generated and journal is rolled to match the bookmark point-in-time, then data integrity and recoverability are improved, but additional processing time and complexity are added
Solution Approach 1:
The system introduces a bookmark as an intermediary element that records the point-in-time state of the second virtual machine. The bookmark serves as a reference marker in the replication journal, allowing the first virtual machine's snapshot to be correlated with the corresponding state at the second site. This intermediary bookmark mechanism enables automatic point-in-time matching without manual intervention or complex synchronization protocols, improving recoverability while minimizing additional processing overhead.
Data Source
AI summary
In one aspect, a method includes pausing write I/Os for a second virtual machine running at a second site, generating a snapshot of a first virtual machine running at a first site after pausing the write I/Os for the second virtual machine, generating a bookmark for the second virtual machine, resuming the write I/Os for the second virtual machine after generating the bookmark, rolling a journal for the second virtual machine to a point-in-time of the bookmark and generating, at the first site, a second snapshot of the second virtual volume for the point-in-time of the bookmark using the journal.