Longitudinal extension method based on super-fusion system and device thereof

Abstract

The invention discloses a longitudinal extension method based on a super-fusion system and a device thereof. The longitudinal extension method comprises the following steps: identifying user configuration so as to match a longitudinal extension storage resource, configuring a first storage layer and a second storage layer; migrating related data stored in the first storage layer and meeting a preset condition to the second storage layer to store, and updating an access address of the migrated related data; acquiring an association coefficient between the migrated related data in the second storage layer and the unmigrated related remained in the first storage layer, acquiring an address sorting value of the association coefficient and a fusion value of the access frequency of the unmigrated related data, if the fusion value is greater than a set value, inserting the migrated related data into the next location behind the storage location of the unmigrated related data, and updating theaccess address of the migrated related data. A read-write operation on continuous partitions can greatly reduce the file fragmentations in the disk, and can short the reading time of the file and improves the system performance at the same time.

Description

technical field [0001] The present invention relates to the technical field of server storage, and more specifically, it relates to a hyper-converged system-based vertical expansion method and a device thereof. Background technique [0002] Hyper-converged infrastructure, also known as hyper-converged architecture, means that the same unit device (x86 server) not only has resources and technologies such as computing, network, storage and server virtualization, but also includes cache acceleration and deduplication. , online data compression, backup software, snapshot technology and other elements, while multiple nodes can be aggregated through the network to achieve modular seamless horizontal expansion and form a unified resource pool. The hyper-converged system is a system based on a hyper-converged architecture, and its horizontal expansion capability is very strong. As long as new physical storage nodes are added to the cluster, the expansion of computing resources and s...

AI Technical Summary

Problems solved by technology

[0003] Generally, after vertical expansion of storage resources in a commercial hyper-converged system, since the storage resources of the nodes used for vertical expansion in the hyper-converged system are much more than other non-vertical expansion nodes, the vertical expansion nodes can provide The storage capacity is also much larger than that of other non-scale-up nodes, but because the hyper-converged system does not prioritize storing data in the storage resources under the scale-up nodes based on the above characteristics, it will divide the access frequency according to the access frequency Low data is stored in the storage resources under the vertical expansion node, but the original storage resources will generate a large number of file fragments due to high frequency access. File fragments refer to discontinuous files generated during the hard disk read and write process

Files written non-sequentially on the hard disk will generate file fragments, which will prolong the seek time of the hard disk and affect system performance

For example, if the virtual memory uses a hard disk, file fragmentation will occur on the hard disk, so disk fragmentation will prolong the seek time of the hard disk and affect system performance.

Images