Full backup method, device and computer-readable storage medium of distributed database system

Abstract

The present invention provides a full backup method, device, and computer-readable storage medium of a distributed database system, wherein the distributed database system includes a plurality of data nodes, and the method includes: acquiring full backup data of each data node of the distributed database system, and the full backup data The full backup time point corresponding to the backup data is the first time point; determine the trust divergence time point at the first time point; obtain the first increment between the trust divergence time point and the first time point of each data node at the first time point Incremental log collection; Obtain the pending transactions of each data node that are in the preparation state and have not been submitted or rolled back from the first incremental log collection; Play back the pending transactions of each data node on the full backup data of each data node . Using the above method, when performing a full backup of each data node of the distributed database system, the branch transactions that are currently in a preparatory state but have not been committed or rolled back can be backed up and stored in the database.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to a full backup method, device and computer-readable storage medium of a distributed database system. Background technique [0002] This section is intended to provide a background or context for implementations of the invention that are recited in the claims. The descriptions herein are not admitted to be prior art by inclusion in this section. [0003] Distributed databases will distribute and store data in multiple nodes (generally called data nodes), figure 1 As a distributed database system, the distributed database system 100 includes a coordinator node 101, multiple data nodes 102, and multiple clients. Wherein, the coordinator node 101 (Coordinator) of the distributed transaction processes the logic as follows: if the global transaction only involves one data node, then the data node can only use the one-phase protocol to submit the transaction. Of course, ...

AI Technical Summary

Problems solved by technology

Of course, it is also possible to use the two-phase protocol for transaction submission, but it will lead to poor processing performance and increase the probability of failure; if the global transaction involves multiple data nodes, then the two-phase (2PC) protocol will be used to start the branch transaction at the data node (also The three-phase protocol can be used), and the coordinator controls the branch transaction to enter the prepare (Prepare) state. After the branch transactions are in the prepare (Prepare) state, a global transaction log (including the branch transaction information corresponding to the global transaction) is recorded. If The global transaction log record is successful (that is, it is considered that the global transaction must be successful), then start to submit the branch transaction in the prepare state until the submission succeeds, if the global transaction log record fails (that is, it is considered that the global transaction will eventually fail), then rollback all branch transactions

[0004] However, when each data node performs a full backup at a certain point in time, the branch transactions in the prepare state are generally not recorded, that is to say, if the full backup time point is in the preparation log of a branch transaction and submitted Before the log lands, the branch transaction cannot be backed up

Images