Search

CN-116126240-B - Storage cluster synchronization method, device and medium

CN116126240BCN 116126240 BCN116126240 BCN 116126240BCN-116126240-B

Abstract

The application discloses a storage cluster synchronization method, a device and a medium, which relate to the field of storage, and in order to improve the replication efficiency in the storage cluster backup, the application controls a storage pool of a main cluster to package information to be synchronized and generate a multiplexed data packet to be synchronized; and controlling the subset group to receive and store the information to be synchronized in the data packet to be synchronized. And sending the data packets to be synchronized to the corresponding sub-clusters by using a multiplexing mechanism, wherein the sub-clusters can timely obtain configuration information of the main clusters and storage information of storage pools in the main clusters and store the configuration information and the storage information to each sub-cluster. Multiple storage pools are transmitted simultaneously by using a multiplexing mechanism, so that the transmission rate is optimized by effectively utilizing the network bandwidth and the copying delay is reduced.

Inventors

  • Lin Shutong

Assignees

  • 济南浪潮数据技术有限公司

Dates

Publication Date
20260512
Application Date
20230106

Claims (9)

  1. 1. A storage cluster synchronization method, comprising: The storage pool of the main cluster is controlled to encapsulate the information to be synchronized, and a multiplexed data packet to be synchronized is generated; Transmitting the data packets to be synchronized to corresponding subsets according to multiplexing and demultiplexing transmission mechanisms; controlling the subset group to receive and store the information to be synchronized in the data packet to be synchronized; the storage pool of the control main cluster encapsulates the information to be synchronized to generate a multiplexed data packet to be synchronized, which includes: controlling the storage pool of the main cluster to generate a snapshot data file and a snapshot configuration file; Comparing the snapshot data file of this time with the snapshot data file of last time; Screening out change information; packaging the change information to be used as the data packet to be synchronized; and adding a unique process identifier and a port number for the encapsulated data packet to be synchronized.
  2. 2. The storage cluster synchronization method of claim 1, wherein the transmission between the primary cluster and the subset cluster is based on a user datagram protocol.
  3. 3. The storage cluster synchronization method according to claim 2, wherein the user datagram protocol is a data transmission protocol based on the user datagram protocol.
  4. 4. A storage cluster synchronization method according to claim 3, wherein said transmitting said data packets to be synchronized to corresponding subsets according to a multiplexing and demultiplexing transmission scheme comprises: Transmitting multiplexed data packets to be synchronized based on the data transmission protocol; Demultiplexing the data packet to be synchronized; and sending the decomposed data packets to be synchronized to the corresponding sub-groups.
  5. 5. The storage cluster synchronization method according to claim 4, wherein after the demultiplexing the data packet to be synchronized, further comprising: judging whether the decomposed data packet to be synchronized is complete or not; If yes, the step of sending the decomposed data packet to be synchronized to the corresponding subset group is entered.
  6. 6. The storage cluster synchronization method according to any one of claims 1 to 5, wherein after said controlling the subset group to receive and store the information to be synchronized in the data packet to be synchronized, further comprises: Judging whether the main cluster is online or not; and if not, controlling the sub-clusters to be online.
  7. 7. A storage cluster synchronization apparatus, comprising: the packaging module is used for controlling the storage pool of the main cluster to package the information to be synchronized and generating a multiplexed data packet to be synchronized; The sending module is used for sending the data packet to be synchronized to the corresponding subset group according to a multiplexing and demultiplexing transmission mechanism; the storage module is used for controlling the subset group to receive and store the information to be synchronized in the data packet to be synchronized; the storage pool of the control main cluster encapsulates the information to be synchronized to generate a multiplexed data packet to be synchronized, which includes: controlling the storage pool of the main cluster to generate a snapshot data file and a snapshot configuration file; Comparing the snapshot data file of this time with the snapshot data file of last time; Screening out change information; packaging the change information to be used as the data packet to be synchronized; and adding a unique process identifier and a port number for the encapsulated data packet to be synchronized.
  8. 8. A storage cluster synchronization apparatus, comprising: A memory for storing a computer program; Processor for implementing the steps of the storage cluster synchronization method according to any of claims 1 to 6 when executing said computer program.
  9. 9. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the storage cluster synchronization method according to any of claims 1 to 6.

Description

Storage cluster synchronization method, device and medium Technical Field The present application relates to the field of storage, and in particular, to a storage cluster synchronization method, apparatus, and medium. Background In the cloud computing era, when cloud loading has become a main architecture scheme of most enterprises, if a server is down, business stop is caused, and the chain reaction and loss caused by the business stop are immeasurable. The high availability of system performance and the early taking of disaster recovery measures have become a normalized knowledge of internet enterprises. Therefore, whether the cloud performance is high or not and whether disaster recovery measures are perfect are becoming one of important standards for an enterprise to measure a public cloud manufacturer. Storage is one of the cores of the internet technology infrastructure, and its highly available solutions are also evolving. The essence of the storage high availability scheme is to copy data to a plurality of storage devices, and the high availability is realized by a data redundancy mode, and the complexity is mainly reflected in how to cope with the problem of data inconsistency caused by copy delay and interruption. The storage clusters in the same area can accelerate data synchronization and reduce inconsistency by improving the data throughput of the storage backboard. After the cloud platform is subjected to cross-regional storage clustering deployment, the master and slave data centers generally ensure data synchronization by building a special network. Although the bandwidth of the cross-regional private network is continuously increasing, the rate of the cross-regional private network is still quite different from that of the local network. In addition, as network Bandwidth Delay Products (BDPs) increase, common methods of using transmission control protocol (Transmission Control Protocol, TCP) transmission protocols begin to become inefficient. The Round Trip Time (RTT) algorithm of the standard TCP has difficulty in achieving high efficiency and fairness in a high BDP environment, and the AIMD algorithm cannot recover the available bandwidth quickly. AIMD is known in English by ADDITIVE INCREASE Multiplicative Decrease. In the transmission control protocol/internet protocol (Transmission Control Protocol/Internet Protocol, TCP/IP) model, belonging to the transport layer, one method for solving congestion control is to increase additively, decrease multiplicatively, or "sum-of-product increase, decrease multiplicatively. Therefore, the provision of a storage high availability scheme for improving replication efficiency is a technical problem to be solved urgently by the person skilled in the art. Disclosure of Invention The application aims to provide a storage cluster synchronization method, a storage cluster synchronization device and a storage cluster synchronization medium for improving replication efficiency. In order to solve the above technical problems, the present application provides a storage cluster synchronization method, including: The storage pool of the main cluster is controlled to encapsulate the information to be synchronized, and a multiplexed data packet to be synchronized is generated; transmitting the data packets to be synchronized to the corresponding sub-groups according to the multiplexing and demultiplexing transmission mechanism; And the control subset group receives and stores the information to be synchronized in the data packet to be synchronized. Preferably, in the above storage cluster synchronization method, controlling a storage pool of a main cluster to encapsulate information to be synchronized, generating a multiplexed data packet to be synchronized includes: Controlling a storage pool of the main cluster to generate a snapshot data file and a snapshot configuration file; Comparing the current snapshot data file with the last snapshot data file; Screening out change information; Packaging the change information to be used as a data packet to be synchronized; a unique process identifier and port number are added to the encapsulated data packets to be synchronized. Preferably, in the storage cluster synchronization method, transmission between the main cluster and the sub-cluster is based on a user datagram protocol. Preferably, in the storage cluster synchronization method, the user datagram protocol is a data transmission protocol based on the user datagram protocol. Preferably, in the storage cluster synchronization method, the sending the data packet to be synchronized to the corresponding subset group according to the multiplexing and demultiplexing transmission mechanism includes: transmitting multiplexed data packets to be synchronized based on a data transmission protocol; Demultiplexing the data packets to be synchronized; and sending the decomposed data packets to be synchronized to the corresponding sub-groups. Preferably, in the above storage c