Snapshot rollback method and device, equipment and storage medium

Abstract

The invention discloses a snapshot rollback method which comprises the following steps: when a rollback instruction is received, recovering a current object state table of a logical volume according to the rollback instruction to obtain an original object state table; performing comparative analysis on the current object state table and the original object state table to determine a change objectin the logical volume; initiating a rollback request about the change object to a storage layer, so that the storage layer rolls back the data in the change object according to the rollback request; according to the snapshot rollback method, the snapshot rollback efficiency is greatly improved. The invention further discloses a snapshot rollback device, snapshot rollback equipment and a computer readable storage medium, which all have the above beneficial effects.

Description

technical field [0001] The present application relates to the field of storage technology, in particular to a snapshot rollback method, and also to a snapshot rollback device, device, and computer-readable storage medium. Background technique [0002] In the distributed block storage service, a logical volume is composed of several fixed-sized objects, and business data is stored in each object. The logical volume uses an object map (object state table) to dynamically manage the state of each object, wherein the object state is mainly divided into non-existing and existing states. [0003] Generally, the process of using the snapshot to perform data rollback on the logical volume data includes restoring the object map corresponding to the original logical volume at the time of the snapshot, and restoring the object data in the original logical volume at the time of the snapshot. However, in the prior art, the object data recovery part is to send all object recovery requests...

AI Technical Summary

Problems solved by technology

However, in the prior art, the object data recovery part is to send all object recovery requests to the storage layer to realize snapshot rollback by the storage layer. However, when the logical volume is large, the number of objects therein is also very large. Large, which leads to a lot of requests sent to the storage layer, and most of them are invalid requests. For example, when the number of objects in the logical volume is tens of millions, the original logical volume has no data change or a small change after the snapshot is created ( That is, only a few object data changes), logically only need to roll back the changed part of the object, but the existing process needs to roll back all tens of millions of objects, resulting in a long rollback time phenomenon

Images