1. Create a single dump file with pg_dump or pg_dumpall. This uses non-parallel PostgreSQL ways to backup the data. Use this option to migrate your data to another database vendor’s system. If restoring to a PostgreSQL or Greenplum database and the dump file is in archive format, you can use pg_restore. If the dump file is in plain text format, you can use a client such as psql. To restore to another Greenplum Database system, do a parallel dump using gpcrondump or gp_dump, then do a non-parallel restore.
2. Create a dump file for each segment with gp_dump. Use this option to backup a database or to migrate your data to a system with the same segment configuration. You must use the gp_restore utility to restore the database. You can use dump files created by gp_dump to restore to a different Greenplum system configuration.
3. Schedule or run routine dumps with gpcrondump. gpcrondump allows you to schedule routine backups, including incremental backups, using the UNIX scheduling utility, cron. Schedule cron jobs that call gpcrondump on the Greenplum master host. gpcrondump backs up databases, data, and objects such as database roles and server configuration files. Full backup jobs scheduled or run with gpcrondump can use Data Domain Boost.
Greenplum DBA FAQ >