Using Parallel Jobs to Dump the PostgreSQL Database

You can use parallel jobs to dump the databases for DumpBasedBackupset subclients. To back up the database, parallel jobs use the pg_dump -j option.

Note

The pg_dump parallel job option (pg_dump with -j) is supported for PostgreSQL 9.3 and more recent versions.
If you schedule a backup on a standby node and the PostgreSQL version is older than 10, do not use parallel jobs.

For PostgreSQL agents, the Indexing version 2 is enabled by default for newly installed clients in Feature Release 11.20 and more recent versions. For clients that are installed in older Feature Releases, the Indexing version will be 1. If you enable the Use parallel jobs to dump the database option for DumpBased BackupSet, the index will be upgraded to version 2.

Before You Begin

Verify that no jobs are running for the instance.
If you have enabled IntelliSnap or block-level backups for an instance, verify that the backup copy jobs are completed.
During pg_dump job, the maximum backup dump size of a database may be up to 120% of the size of the database. Ensure that the staging path has sufficient free space to accommodate subclient content and the number of streams.

Procedure

From the CommCell Browser, go to Client Computers > client > PostgreSQL > instance.
Right-click the DumpBasedBackupSet, and then click Properties.

The Backup Set Properties dialog box appears.
Select the Use parallel jobs to dump the database check box.
In the Number of parallel jobs box, enter the number of jobs to dump the database.

Note

The pg_dump -j option uses multiple database connections to connect to the database, once using the leader process and one more time for each worker job. As pg_dump opens n jobs+1 connections to the database per stream, verify that the max_connections setting is set to the sufficient value that can accommodate all connections.
To set the staging path for backups, in the Staging Path box, enter the location to use for staging backups.
Click OK.