Oracle RMAN Plugin retention issue

[rleon]

Hi all,

I am currently testing the Oracle RMAN Plugin for an RAC Cluster. Backups and restores are working well.
For deleting old backups from the Veeam Server Repository, I have:

RMAN>CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 14 DAYS;

And then I also periodically run a script that deletes obsolete RMAN backups from the Veeam Server Repository:

RMAN>DELETE OBSOLETE;

This works fine. RMAN backups that are older than 14 days are deleted from the Repository.
I don't even have to configure the Plugin's "OracleRMANConfigTool --set-force-delete".
Side note: I do not recommend using "OracleRMANConfigTool.exe --set-force-delete" because the RMAN/Recovery Catalog records are not updated, which means you will eventually have to run RMAN CROSSCHECK.

However, I can see a potential issue here.
What if the Oracle database server goes offline for an extended period, then the Veeam Server Repository backups will not be deleted, ever.
That is because both "RMAN>DELETE OBSOLETE" and "OracleRMANConfigTool --set-force-delete" are directed by the Oracle server, and not by the Veeam Server itself.
Sure, I can go to the Veeam Console and right-click on the RMAN backups to delete them, but currently, there is no way to delete specific older unwanted RMAN backups. You can only delete ALL RMAN backups for an Oracle server.

I just wish there is a way for the Veeam server to set a retention period for RMAN Plugin backups and delete them to release disk space, without the Oracle server having to be online.
Sure, you'd then have to run RMAN CROSSCHECK later, but still, the issue here is that the releasing of Repository disk space is currently dependent on the Oracle server being online.
This could be a little wasteful on the Veeam Repository disk space in some situations.
But overall, I suppose this is a minor issue.

[HannesK]

Hello,
thanks for the request. As far as I remember, the Veeam Backup server does not have any knowledge about the retention configured in RMAN. So "it's complicated".

In general: also for VM and agent jobs, we require job runs to apply retention. In V11, we will apply retention for GFS restore points even without job run. But for normal backups in V11, they still stay until a job runs or an administrator manually deletes.

Best regards,
Hannes

[rleon]

Hi HannesK,

Not sure about v11 until I get to test it.
My problem is not that I cannot manually delete RMAN backups from the Veeam Console, but that the Console-delete-route current lacks 2 things:
1. Delete a specific RMAN backup for an Oracle server instead of deleting ALL backups. (I understand this is already true for "normal" backups from VM/Agent jobs, where you also cannot delete just a specific backup of a specific day)
2. If I use the Oracle Plugin to delete specific backups, either through RMAN retention policy plus DLEETE OBSOLETE commands, or "OracleRMANConfigTool --set-force-delete"... Sure I can achieve "deleting a specific RMAN backup instead of deleting ALL", but this implies that the the Oracle server itself will need to be online for this whole thing to work. This is not a good dependency to have.
I.e., useless RMAN backups that are no longer needed by the DBA could be stuck wasting precious Veeam Repository disk space if the DBA decides to switch off/decommission their Oracle DBs without properly informing the Veeam admins. I understand this is a bit of a corporate policy and communication issue, but none of the other types of Veeam jobs has this potential problem that the Plugin jobs have.