Method and device for adjusting service speed based on down-refreshing speed

Abstract

The invention provides a method, a system and device for adjusting a business speed based on a down-refreshing speed and a storage medium, and the method comprises the steps: setting an early warning threshold value, and writing the early warning threshold value into a memory; creating a down-brushing thread to brush data in the cache pool down to a storage pool, and detecting whether the current water level of the cache pool exceeds the early warning threshold value or not every preset time; responding to the situation that the current water level of the buffer pool exceeds the early warning threshold value, and applying for a write-in permission according to the current down brushing rate of the buffer pool; and writing the data of the client into the cache pool according to the writing permission. According to the method, when the water level exceeds the early warning threshold value, the permission is applied for the write-in request based on the counted down-refreshing rate, and the actual execution of the service can be executed only when the permission is obtained, so that the cache pool is effectively prevented from being fully written.

Description

technical field [0001] The present invention relates to the field of distributed storage systems, and more specifically refers to a method, system, device and storage medium for adjusting service speed based on download speed. Background technique [0002] In the current era of big data, data collection, transmission and storage are facing many challenges; in the face of massive data, efficient and consistent storage is required, which has become an urgent need for various business systems; The ultimate data storage medium, HDD (Hard Disk Drive, mechanical hard disk) still has a very large advantage in terms of storage capacity and cost performance, but in the face of the upper-level business system's IOPS (Input / Output Per Seconds, input and output per second) volume) has become more and more demanding; in terms of small IO read and write speeds, SSD (Solid State Disk / Drive, solid state drive) / NVMe (Non Volatile Memory Express, non-volatile memory standard) and other new st...

AI Technical Summary

Problems solved by technology

[0002] In the current era of big data, data collection, transmission and storage are facing many challenges; in the face of massive data, efficient and consistent storage is required, which has become an urgent need for various business systems; The ultimate data storage medium, HDD (Hard Disk Drive, mechanical hard disk) still has a very large advantage in terms of storage capacity and cost performance, but in the face of the upper-level business system's IOPS (Input / Output Per Seconds, input and output per second) volume) has become more and more demanding; in terms of small IO read and write speeds, SSD (Solid State Disk / Drive, solid state drive) / NVMe (Non Volatile Memory Express, non-volatile memory standard) and other new storage The medium has obvious advantages, but it cannot be used in large quantities in commercial systems due to the high price and relatively low cost performance.

[0003] How to meet the actual needs of the business system based on the actual situation of the current hardware storage is a practical problem that needs to be solved. When implementing the storage system, it will do hierarchical processing, that is, use a small amount of SSD / NVMe as a cache pool to improve the IOPS of the storage cluster , the backend uses a large number of HDDs for persistent storage of data, and the storage system solidifies the data from the buffer pool to the underlying storage pool through the data flushing strategy; but because the IOPS of SSD / NVMe far exceeds that of HDD, the business continues When writing at full speed, the cache pool will be full, resulting in denial of service for the entire storage cluster

Images