Snapshot method and device for a distributed storage system

Abstract

An embodiment of the invention discloses a snapshooting method of a distributed storage system. The snapshooting method includes: firstly, snapshooting a base image of the distributed storage system CEPH to acquire a snapshot image of the current base image; then cloning the snapshot image and setting a cloned image to be a current operation image used for executing read-write operation of subsequent data. When the CEPH is subjected to snapshooting, an ordered operation mode in snapshooting and cloning of RBD (rados block device) is adopted, the subsequent data is written to the cloned image directly each time it is written after creation of snapshot, only one cloned-image writing operation is involved, operations in reading the base image once and writing the snapshot once are omitted, reduction of disk pressure is benefited, write-operation performance in snapshooting the CEPH is improved, read-operation performance is improved as well, and user experience of a cloud platform is improved. The invention further provides a corresponding implementation device which enables the method to be practical and has the corresponding advantages.

Description

technical field [0001] Embodiments of the present invention relate to the field of storage technologies, and in particular, to a snapshot method and device for a distributed storage system. Background technique [0002] CEPH is a distributed storage system with high scalability, high reliability and high performance, supporting file system, block storage and object storage; with the help of the rapid development of OpenStack and its open source features, it has attracted more and more developers Practice and explore in a production environment. Relying on the unified storage system it provides, CEPH is widely used in the construction of cloud data centers, where block storage is deployed by production systems due to its stability. CEPH provides two usages of calling the librbd library through the kernel module or user mode to use its block storage. The block device in CEPH is represented as an RBD (Rados Block Device) image, where rbd is the user of the Linux driver and QEM...

AI Technical Summary

Problems solved by technology

[0006] It can be seen from the above that after the snapshot operation is created by CEPH, the next data write operation may cause the problem of reading once and writing twice. For scenarios with frequent write operations, a large amount of data replication will lead to business collapse. ;In addition, the recovery operation will cause the full amount of snapshot data at the recovery time to be copied to the original image file, which will also bring a large amount of data copy, resulting in high disk writing pressure, and in severe cases, it will cause the system to crash

Images