Hyper-fusion system data localization storage method based on node evaluation function

Abstract

The invention relates to a hyper-fusion system data localization storage method based on a node evaluation function, and relates to the technical field of data storage. A centralized distributed storage system is used as a bottom storage device of the system, a mode that a local physical node and a plurality of far-end physical nodes jointly maintain data copies is adopted, a data block of a virtual machine instance in the system stores a main copy at the physical node where the data block is located, data balance distribution between the nodes is considered, the hardware performance and the relative storage utilization rate of the physical nodes are comprehensively considered, the other two far-end physical nodes are selected to store the second data copy and the third data copy, complete localization storage of virtual machine data is guaranteed, and therefore when the virtual machine reads the data, needed data can be directly read from the physical node where the virtual machine is located, the data reading I / O path of the virtual machine is shortened, cross-physical-node reading is avoided, and the data reading delay of the virtual machine is reduced.

Description

technical field [0001] The invention relates to the technical field of data storage, in particular to a localized storage method for hyper-converged system data based on node evaluation functions. Background technique [0002] Decentralized distributed storage systems usually use hash algorithm or its derivative algorithm to calculate the data nodes where multiple data copies are placed. Based on this, the data distribution on the hyper-converged system has significant dispersion and randomness. Virtual The data of the virtual machine is not completely stored on the physical node where the virtual machine is located. For each virtual machine instance in the hyper-converged system, each physical node of the system stores a part of its data. When the virtual machine reads its own data, it needs to frequently copy the required data blocks from other nodes to the virtual machine through the inter-node network The physical node where it is located causes a large network bandwidt...

AI Technical Summary

Problems solved by technology

For each virtual machine instance in the hyper-converged system, each physical node of the system stores a part of its data. When the virtual machine reads its own data, it needs to frequently copy the required data blocks from other nodes to the virtual machine through the inter-node network The physical node where it is located causes a large network bandwidth occupation, and also brings a huge delay in reading data, and this situation is significantly aggravated as the number of physical nodes in the system increases

The centralized distributed storage system has metadata service nodes and data service nodes, so the data service nodes for data storage can be specified, but random selection is used when selecting slave nodes, which can easily lead to data imbalance, and this situation occurs in the system It is more serious when the number of physical nodes is large

Images