Note: I have no experience with Oracle Database administration and I don't perform it. Another team set up the new backup using RMAN and I was troubleshooting the issue on the Veeam side. This was Veeam Backup & Replication 12.3.2.
Introductory Information
I described backing up Oracle Database using Veeam Plug-in for Oracle RMAN in the articles:
Official description of Plug-in configuration in Linux and information for Oracle RAC:
Error During Job Execution
Almost immediately after starting the (new) job, an Error appeared and the job terminated. Existing backups of other databases on the cluster continued running without issues.
20.10.2025 2:56:12 :: Oracle RMAN Backup job started at 20.10.2025 2:56:14 20.10.2025 2:56:12 :: Session failed: Backup is not in user's control scope
Current Environment
There are several physical servers running Oracle Database. Two of these servers are configured as Oracle RAC (Real Application Clusters). Veeam Plug-in for Oracle RMAN running in Standalone Mode requires entering a user account that authenticates to the VBR Server. Our environment uses a unique account for each server.
In the case of Oracle RAC, backing up Cluster Databases is configured only from the first server (although it is recommended to set it on all RAC servers). I later discovered that the error occurred when a new Standalone Database backup was configured on the second server.
Troubleshooting
To resolve the issue, it's good to determine detailed information about when the error occurs. Alternatively, test various situations. A universal approach is to find more detailed logs to see if they contain more information.
Logs on VBR Server
From the documentation we learn that logs on the Windows Veeam Backup Server can be found in the folder
%ProgramData%\Veeam\Backup\Plugin\Backup\
There are subfolders for individual servers, nested twice within each other, and a Session folder where numerous logs are located. For current logs, the Session.log file is of interest.
The log is quite long, even though the job fails right at the beginning. There are several interesting entries that led me in the right direction.
[06:57:26.709] <182> Info (3) [LinuxBackupServerWorker] Command: plugin.findOrUpdateBackupByJobId
[06:57:26.725] <182> Error (3) Async command failed
[06:57:26.725] <182> Error (3) Backup is not in user's control scope (System.Exception)
[06:57:26.725] <182> Error (3) at Veeam.Backup.VmbAPI.CForeignInvokerBackupDbAccessor.GetBackupsByJobIdThrowIfWrongOwner
(Guid jobId)
[06:57:26.725] <182> Error (3) at Veeam.Backup.VmbAPI.CVmbBackupScope.FindByJobIdThrowIfWrongOwner(Guid jobId)
[06:57:26.725] <182> Info (3) [LinuxBackupServerWorker] Result: {(EString) exception=Backup is not in user's control scope;
(EGuid) commandId={4b3deb00-b192-4f98-b284-cb5646e9b45e}; }
Backup Job Names for Oracle RMAN
The Backup Job Name for Veeam Plug-in for Oracle RMAN running in Standalone Mode is created automatically. For a standalone server, the server name (hostname) and repository name are used. It may look like this
oracledb.company.cz Oracle backup (Repository-Name)
For servers in an Oracle RAC environment, the cluster name - Single Client Access Name (SCAN) and repository name are used. The result looks similar, but for all RAC members the job name is the same. This likely has a purpose, so that it doesn't matter which server the cluster backup runs from, and the data is stored together.
rac-cluster-scan.company.cz Oracle backup (Repository-Name)
Job and Its Owner
I found that standard backups (cluster databases) were only running from the first server in the cluster. The Veeam Plug-in wasn't installed on the second one at all. My colleagues set up a standalone database on the second server, installed the plugin, and configured the backup. The error appeared immediately on the first run.
Because the servers are in RAC, the Veeam Plug-in created the same job name. The second server tried to connect to the job created by the first server and apparently doesn't have the rights to do so. According to the error FindByJobIdThrowIfWrongOwner, it is not the owner of the job.
It's quite possible that the problem is caused by the fact that each server uses a different account to connect to the VBR Server. Perhaps a shared account needs to be used within the Oracle RAC. I found no information about this in the Veeam documentation. Nor about whether Standalone Databases are supported when backing up Oracle RAC.
Workaround
According to the documentation, there may be situations where different standalone servers have the same name (in different environments). Therefore, there is an option to modify the configuration and differentiate server names to create unique job names.
The documentation states that the following procedure applies only to Standalone servers. It worked practically in my case with Oracle RAC as well (it may not be correct to use it due to cluster operation, but in our case it works). After setting the custom server name parameter, a job was created where this name was used in its name (instead of the SCAN name).
- the configuration is located on Oracle server in the XML file
veeam_config.xml, which on Linux is stored in the path/opt/veeam/VeeamPluginforOracleRMAN/ - we find the line
<PluginParameters /> - we add the
customServerNameattribute with the desired FQDN name
<PluginParameters customServerName="oracledb2.company.cz" />
Note: By the way, this configuration file contains a lot of sensitive information. Veeam recommends restricting access to it - Granting Permissions to Users.
There are no comments yet.