Identical Snapshots for Stretched Volumes via Write Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for creating identical snapshots of volumes in a stretched volume configuration require quiescing and draining of host I/Os, leading to additional latency and adverse impacts on applications.
Innovation Solution
A method that utilizes a write tracking mechanism, such as a transactional caching layer, to track tagged write I/Os without requiring quiescing and draining, allowing for the creation of identical snapshots of volumes without interrupting I/O operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If quiescing and draining of host I/Os is performed to create identical snapshots, then snapshot consistency is achieved, but additional latency is incurred and application performance deteriorates
Solution Approach 1:
The system performs preliminary tagging of write I/Os with unique identifiers before snapshot creation. This preliminary action enables the tracking and synchronization of writes across replicated volumes during snapshot operations, ensuring consistency without requiring I/O quiescing that would cause latency.
Solution Approach 2:
The patent replaces the mechanical approach of stopping I/O operations (quiescing) with a software-based tracking mechanism using tagged identifiers. Instead of physically interrupting the data flow, the system uses metadata tagging and tracking to achieve snapshot consistency, eliminating the performance penalty associated with I/O suspension.
2Manufacturing precision
If quiescing and draining of host I/Os is performed to create identical snapshots, then snapshot consistency is achieved, but application performance is adversely impacted
Solution Approach 1:
The system introduces an intermediary tracking mechanism that monitors and coordinates write operations across replicated volumes. This intermediary layer uses tagged identifiers to track writes without interfering with the actual data flow, enabling snapshot consistency while maintaining application performance by avoiding direct I/O interruption.
Solution Approach 2:
The patent substitutes the mechanical I/O suspension method with a software-based tracking and coordination mechanism. By using tagged identifiers and metadata tracking, the system achieves snapshot consistency through information management rather than physical I/O control, preserving application performance.
3Loss of time
If write tracking mechanism is used to create identical snapshots, then additional latency is avoided, but system complexity increases
Solution Approach 1:
The system segments the snapshot creation process into distinct phases: tagging write I/Os with identifiers, tracking tagged writes across replicated volumes, and synchronizing based on tracked information. This segmentation allows each component to perform a specific function efficiently, reducing overall system complexity while avoiding latency.
Solution Approach 2:
The patent uses copying of tagged identifier information across the replication network to track writes without duplicating the actual data movement. By copying metadata (tagged identifiers) rather than full data blocks, the system reduces complexity while maintaining snapshot consistency and avoiding latency.
Data Source
AI summary
In at least one embodiment, processing can include: receiving a request to create identical snapshots of volumes V1 and V2 configured as a stretched volume; and in response to receiving the request, performing first processing including: holding write acknowledgements for the stretched volume; tracking writes to the stretched volume; creating a first snapshot of V1 on a first system and a second snapshot of V2 on a second system; stopping tracking of writes to the stretched volume; resuming write acknowledgements for the stretched volume; determining tracked writes for the stretched volume; determining a set of locations corresponding to the tracked writes; selecting the first snapshot of V1 as a master copy; determining data changes corresponding the set of locations from the master copy; replicating the data changes from the first system to the second system including the second system; and applying the data changes to the second snapshot of V2.


