NetVault Plug-in for PostgreSQL 11.1 - User Guide

NetVault Backup 10.x service (netvault-pgsql) does not start on Windows

Check the Windows Event Viewer for the following message: PDT FATAL: lock file “postmaster.pid” already exists.

NetVault Backup 10.x uses a PostgreSQL database. If the PostgreSQL database does not start, NetVault Backup cannot start. To correct this issue, delete the “postmaster.pid” file from the location referenced in the log and restart the NetVault Backup Server. For more information, see https://support.quest.com/netvault-backup/kb/122475.

Login failed

Cannot connect to the database server.

The PostgreSQL Server is not running. Start the services.

Password Authentication Failed for the User <username>.

The superuser password set in the Configure dialog box is invalid. Update the password in the plug-in by following the instructions in Reconfiguring an added database cluster.

Authentication request failed.

The “pg_hba.conf” file is not available for user authentication.

If the “pg_hba.conf” file contains entries for both IPv4 and IPv6, configure the same authentication for each. For example, use TRUST authentication for both IPv4 and IPv6 connections.

Backup failed

Login Failed.

Backup in TAR Archive File format fails with the following message:

Database found, but failed to backup Database <dbname>.

Check the pg_dump command with the TAR Archive format from a command prompt or terminal session. If the following error is displayed, “pg_dump: [tar archiver] actual file length <xxxx> does not match expected <xxxx>,” you cannot use the TAR Archive File backup format on the current installation. Use the Custom Archive File format instead.

On Windows, the backup fails with the following message:

Database was found, but failed to backup <locale_name> database.

Check the pg_dump command from the command prompt. If the following error is displayed, “pg_dump: Cannot access “??-UTF8-C” [database to be dumped]: FATAL: database “??-UTF8-C” does not exist pg_dump: *** Stopped by error,” the problem is in the character-set encoding configured for the database cluster.

The backing up of databases or tables named with non-ASCII characters is not supported if the encoding is set to UTF-8 at initdb.

Appropriate encoding must be set for the database cluster, for example, EUC_JP for Japanese or EUC_KR for Korean, to back up the databases and tables.

Backup fails with the following message:

Database found, but failed to backup Database XXXXX.

Ensure that the password provided in the Add Database Cluster/Configure Database Cluster dialog box of the plug-in is correct. If the cluster is configured to use TRUST authentication for HOST connection while PASSWORD authentication for LOCAL connection, the backups fail if the password is incorrect or blank.

If the “pg_hba.conf” file contains entries for both IPv4 and IPv6, configure the same authentication for each. For example, use TRUST authentication for both IPv4 and IPv6 connections.

You can browse the tree and submit a job with a blank or incorrect password. However, the actual job fails.

Quest recommends the following in this scenario:

Backup Completed with warnings

If the backup job shows this status, check the NetVault Backup Log entries for the job to see if they display one or both of the following messages:

These messages indicate that the selected data was backed up, but the job’s index information was not properly added by NetVault Backup to its database.

Without this index information, the data cannot be properly restored.

Method 1:

Access the Manage Devices page of the NetVault Backup WebUI, and perform a scan of the media targeted by the job. NetVault Backup stores index information for backup jobs in two locations: in the NetVault Database and on the media targeted by the backup. Performing this scan adds the index information to the NetVault Database. To verify that the information was added, open the Create Restore Job — Choose Saveset page and locate the specific job. If you can browse it and set up a restore job, the scan process has corrected the problem.

Method 2:

If the previous method failed, run the backup job again.

Restore failed

Login Failed.

Restore job failed and PostgreSQL log lists the following message:

ERROR: cannot drop constraint orders_pkey on table orders because other objects depend on it. HINT: Use DROP … CASCADE to drop the dependent objects too.

Your PostgreSQL database uses foreign keys to maintain the integrity of the data.

To drop the table manually from PostgreSQL, use the DROP TABLE command with the CASCADE parameter. DROP TABLE removes the table from the database, and CASCADE ensures that any objects that depend on the table are also dropped. For more information, see your PostgreSQL documentation.