Efficient parallel database backup and recovery tool

table of Contents

 

一 、 gpbackup / gprestore

Two, gpcopy


一 、 gpbackup / gprestore

Starting from version 5.5.0, Greenplum Database has released a more efficient Greenplum-based gpbackup/gprestore utility based on the built-in COPY...ON SEGMENT command. For a detailed introduction to the COPY……ON SEGMENT command, please refer to section 6.1.1. In general, gpbackup only stores metatable files and DDL files of objects, and the generation, compression, and storage of backup files are completed on each segment, so it is more efficient. The meta table information of gpbackup contains all the information needed for gprestore to run. In addition, storing the data in csv format allows the data to be loaded into the same cluster or another cluster by other tools (such as gpload). Each gpbackup task uses a single Greenplum transaction. During the execution of the transaction, the meta table information will be backed up to the Master node, and the data files will be backed up to the Segment node in parallel through the COPY...ON SEGMENT command. During the backup process, the backup process will only obtain the ACCESS SHARE table lock on the backup table, and will not block the online Greeplum from providing services to the outside world.

Specifically&#x

Guess you like

Origin blog.csdn.net/MyySophia/article/details/113991267