Storage System Elastic Space for Host Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems where the production host has a larger cache SSD capacity than the backup host, failover processes are hindered by the backup host's limited caching capacity, leading to low readability and extended failover times due to data being transferred from slow disks after being considered 'cold' by the storage system.
Innovation Solution
Establishing an elastic space in the storage system's fast disk to store metadata of data blocks, allowing the expansion of storage capacity and dynamic migration of data blocks to improve failover efficiency by ensuring data requested by the backup host is retrieved from fast disks during failovers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the production host is configured with a larger cache SSD capacity than the backup host, then the production host can cache more data blocks locally, but the backup host's limited caching capacity causes longer failover times when data must be retrieved from slow disks
Solution Approach 1:
The storage system proactively identifies data blocks that the production host has cached and pre-caches them in the elastic space before failover occurs. This preliminary action ensures that when failover happens, the backup host can immediately access these data blocks from the fast disk elastic space rather than retrieving them from slow disks after the fact, thus resolving the contradiction between asymmetric cache capacities and failover time
Solution Approach 2:
The elastic space acts as an intermediary buffer between the production host's large cache and the backup host's small cache. It stores copies of data blocks that the production host has cached, enabling the backup host to access this data through the fast disk elastic space during failover, effectively bridging the capacity gap between the two hosts' cache SSDs
2Productivity
If data blocks are transferred to slow disks after being cached by the production host, then storage space is optimized, but readability decreases and failover time increases when the backup host needs this data
Solution Approach 1:
The system performs preliminary actions by identifying and pre-caching data blocks in the elastic space before failover occurs. This ensures that data remains in fast disk storage and is not transferred to slow disks when the backup host needs it, maintaining both storage efficiency and data readability during critical failover operations
Solution Approach 2:
The elastic space provides a localized fast storage solution specifically for data blocks that are likely to be needed during failover. This local quality enhancement ensures that critical data remains highly readable in the fast disk elastic space while other less critical data can reside on slow disks, thus resolving the contradiction between storage efficiency and data readability
Data Source
AI summary
Performing failover processing between a production host and a backup host, a storage system is connected to the production host and the backup host. In response to a failure of the production host, metadata is obtained of data blocks that have been cached from an elastic space located in a fast disk of the storage system. A storage capacity of the elastic space is expanded. Data blocks are obtained to which the metadata corresponds according to the metadata and the storage capacity of the expanded elastic space, and storing the same in the expanded elastic space. In response the backup host requesting the data blocks to which the metadata corresponds, and the data blocks to which the metadata corresponds have already been stored in the expanded elastic space, data blocks are obtained to which the metadata corresponds from the expanded elastic space and transmitting the same to the backup host.