Backing Up with gpcrondump
A newer version of this documentation is available. Click here to view the most up-to-date release of the Greenplum 4.x documentation.
Backing Up with gpcrondump
Use gpcrondump to backup databases, data, and objects such as database roles and server configuration files.
The gpcrondump utility dumps the contents of a Greenplum database to SQL script files on the master and each segment. The script files can then be used to restore the database.
The master backup files contain SQL commands to create the database schema. The segment data dump files contain SQL statements to load the data into the tables. The segment dump files are compressed using gzip. Optionally, the server configuration files postgresql.conf, pg_ident.conf, and pg_hba.conf and global data such as roles and tablespaces can be included in a backup.
gpcrondump -x mydbThis performs a full backup of the specified database to the default locations.
gpcrondump mydb -u /backupsThe gpcrondump utility creates the db_dumps subdirectory in the specified directory. If there is more than one primary segment per host, all of the segments on the host write their backup files to the same directory. This differs from the default, where each segment writes backups to its own data directory. This can be used to consolidate backups to a single directory or mounted storage device.
In the db_dumps directory, backups are saved to a directory in the format YYYYMMDD, for example data_directory/db_dumps/20151012 for a backup created on October 12, 2015. The backup file names in the directory contain a full timestamp for the backup in the format YYYYMMDDHHMMSS, for example gp_dump_0_2_20151012195916.gz. The gpdbrestore command uses the most recent backup by default but you can specify an earlier backup to restore.
If you include the -g option, gpcrondump saves the configuration files with the backup. These configuration files are dumped in the master or segment data directory to db_dumps/YYYYMMDD/config_files_timestamp.tar. If --ddboost is specified, the backup is located on the default storage unit in the directory specified by --ddboost-backupdir when the Data Domain Boost credentials were set. The -G option backs up global objects such as roles and tablespaces to a file in the master backup directory named gp_global_-1_1_timestamp.
If --ddboost is specified, the backup is located on the default storage unit in the directory specified by --ddboost-backupdir when the Data Domain Boost credentials were set.
There are many more gpcrondump options available to configure backups. Refer to the Greenplum Utility Reference Guide for information about all of the available options. See Backing Up Databases with Data Domain Boost for details of backing up with Data Domain Boost.
Backing up a database with gpcrondump while simultaneously running DDL commands might cause issues with locks. You might see either the DDL command or gpcrondump waiting to acquire locks.
prefix_gp_dump_content_dbid_timestampThe content and dbid are identifiers for the Greenplum Database segment instances that are assigned by Greenplum Database. For information about the identifiers, see the Greenplum Database system catalog table gp_segment_configuration in the Greenplum Database Reference Guide.
prefix_gp_dump_[0 or 1]_dbid_timestamp
Where the value 0 is for segment instances and the value 1 is for the master instance.
The gpdbrestore utility in Greenplum Database 22.214.171.124 and later releases recognizes both backup file name formats.