Storage Partition Extension via Index Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face challenges in extending partitions without causing downtime, leading to service interruptions when storage space is insufficient, as current methods require stopping storage system services to re-divide partitions.
Innovation Solution
A method that divides existing partitions into two new partitions, generating corresponding partition indexes through replication and copy-on-write technology, allowing the system to manage objects efficiently and process access requests without downtime by utilizing the original partition index.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If partition extension is performed by re-dividing partitions, then storage capacity is improved, but service interruption occurs
Solution Approach 1:
The patent creates a new partition table structure in advance before the extension operation. The new partition table is prepared with the correct configuration for the extended storage capacity, and then the system switches from the old partition table to the new one without interrupting services. This preliminary preparation allows the extension to occur seamlessly.
Solution Approach 2:
The patent creates a copy of the partition table structure during the extension process. A new partition table is generated as a copy of the existing one, modified to reflect the extended capacity, and then activated. This copying mechanism allows the system to maintain the old partition table as a backup while using the new one for extended storage.
2Quantity of substance
If partition extension is performed by re-dividing partitions, then storage capacity is improved, but downtime is caused
Solution Approach 1:
The patent ensures continuous service operation during the partition extension process. By using a partition table switching mechanism where the new partition table is activated without stopping services, the system maintains continuous useful action. The old partition table remains valid until the switch is complete, ensuring no interruption in service.
Solution Approach 2:
The new partition table is prepared in advance before the extension is activated. This preliminary action includes creating the partition table structure, configuring it for the extended capacity, and validating it. Once prepared, the switch to the new partition table can occur quickly without requiring service interruption, thus minimizing downtime.
3Quantity of substance
If partition extension is performed by re-dividing partitions, then storage capacity is improved, but system complexity increases
Solution Approach 1:
The patent uses a copying mechanism where the new partition table is created as a copy of the existing partition table structure. This approach simplifies the extension process because the basic structure is already defined and validated. The system only needs to modify the copy with the new capacity parameters rather than creating an entirely new complex structure.
Solution Approach 2:
The partition table switching mechanism is established in advance as part of the system architecture. By having the switching capability pre-configured, the actual extension operation becomes a simpler process of activating the pre-prepared partition table rather than implementing a complex real-time reconfiguration during the extension.
Data Source
AI summary
A partition of a storage system is extended in a more effective manner, thereby improving the overall operating efficiency of the storage system. For instance, objects are divided into a first group of objects and a second group of objects, based upon which a first partition including the first group of objects and a second partition including the second group of objects are created. A partition index of the partition is obtained, wherein a group of nodes included in the partition index respectively represents metadata of the group of objects. A first partition index and a second partition index are respectively generated based on the partition index, wherein a first group of nodes included in the first partition index respectively represents metadata of the first group of objects, and a second group of nodes included in the second partition index respectively represents metadata of the first group of objects.