Skip to main content

Backup and Restore Overview

Cloudberry Database offers both parallel and non-parallel methods for database backups and restores. Parallel operations handle large systems efficiently because each segment host writes data to its local disk at the same time. Non-parallel operations, however, transfer all data over the network to the coordinator, which then writes it to its storage. This method not only concentrates I/O on a single host but also requires the coordinator to have enough local disk space for the entire database.

Parallel backup with gpbackup and gprestore

Cloudberry Database provides gpbackup and gprestore for parallel backup and restore utilities. gpbackup uses table-level ACCESS SHARE locks instead of EXCLUSIVE locks on the pg_class catalog table. This enables you to execute DDL statements such as CREATE, ALTER, DROP, and TRUNCATE during backups, as long as these statements do not target the current backup set.

Backup files created with gpbackup are designed to provide future capabilities for restoring individual database objects along with their dependencies, such as functions and required user-defined data types.

For details about backup and restore using gpbackup and gprestore, see Perform Full Backup and Restore and Perform Incremental Backup and Restore.

Non-parallel backup with pg_dump

You can also use the PostgreSQL non-parallel backup utilitiesmpg_dump and pg_dumpall to create a single dump file on the coordinator host that contains all data from all active segments.

The PostgreSQL non-parallel utilities should be used only for special cases. They are much slower than using gpbackup and gprestore because all of the data must pass through the coordinator. In addition, it is often the case that the coordinator host has insufficient disk space to save a backup of an entire distributed Cloudberry Database.

The pg_restore utility requires compressed dump files created by pg_dump or pg_dumpall. Before starting the restore, you should modify the CREATE TABLE statements in the dump files to include the Cloudberry Database DISTRIBUTED clause. If you do not include the DISTRIBUTED clause, Cloudberry Database assigns default values, which might not be optimal.

To perform a non-parallel restore using parallel backup files, you can copy the backup files from each segment host to the coordinator host, and then load them through the coordinator.

Another non-parallel method for backing up Cloudberry Database data is to use the COPY TO SQL command to copy all or a portion of a table out of the database to a delimited text file on the coordinator host.