How Publishing Works
Publishing PostgreSQL instances with Veeam Explorer for PostgreSQL works in the following manner:
- To start the publishing process, Veeam Explorer for PostgreSQL sends a publishing command to the Veeam Mount Service running on the mount server associated with the backup repository.
- The Veeam Mount Service delegates this request to the Veeam Explorers Recovery Service running on the same server.
- The Veeam Explorers Recovery Service connects to the target server. The service validates the permissions of the selected user and checks if there is enough free space on the target server. The Veeam Explorers Recovery Service sends a request to the Veeam Mount Service to connect to the backup repository and initiate the mounting operation.
- The Veeam Explorers Recovery Service uses FUSE to mount the necessary file system to the /tmp directory on the target PostgreSQL machine. The service opens the instance from the mounted file system, so that you can perform the required operations with PostgreSQL tools.
After you have launched a publishing operation to a PostgreSQL server, you can quickly republish the latest or point-in-time state of the PostgreSQL instance to the same server.
The publishing session is resilient to network disruption, backup server or mount server crashes. All changes in instance files that occur after publishing are saved in the publishing write cache, which is stored in the /var/tmp directory on the target server. If anything disrupts the publishing process (the target or mount server crashes, or the network is down), you can launch the retry manually after the server or network is up.
Starting from Veeam Backup & Replication 12.1, you can export the modified databases managed by the published instance. For more information, see Exporting From Published Instances.