I'm performing a proof of concept regarding the rman backup to a scale-out repository (composed of a performance tier in a windows host, and capacity tier of S3 storage)
The setup was done successfully and the backup are going properly to the performance tier
However, i'm not getting any backups copied from the performance tier to the capacity tier while i ticked the box to "copy backups as soon as they are created" box
the SOBR job is starting but it finds 0 backups to copy while backups are present in my performance tier
Is there some limitation regarding the rman backups copy to a capacity tier (S3) ? or it can be a technical issue?
Have you waited for 24 hours? The backup policy uses the same backup file (*.VAB) for a 24-hour period. It can be copied or moved to the capacity tier after a new backup file has been created.
Important
With both the Copy and Move policies, Veeam Backup & Replication copies and moves to the capacity tier only non-active backup files created more than 24 hours ago, which match the capacity tier configuration parameters. For details, see Backup Chain. To learn more about the capacity tier configuration, see Add Capacity Tier in the Veeam Backup & Replication User Guide.
"With the Copy policy, Veeam Backup & Replication checks backup files every 1 hour and runs an offload job only if new backup files exist."
but it is also written just after (my bad!)
With both the Copy and Move policies, Veeam Backup & Replication copies and moves to the capacity tier only non-active backup files created more than 24 hours ago, which match the capacity tier configuration parameters.
I'm not sure what is meaning of the "non-active" files. maybe it means outside of the recovery window? i hope no
In any case it means that for RMAN backups i cannot considered the capacity tier as a secondary backup target because of that delay of 24h (RPO for secondary copy > 24H)
We use a single backup file ("container") to store backups produced by multiple backup sessions. It allows us to improve overall performance and scalability by reducing the number of objects on a repository file system and in our configuration database. By default, we keep this file open for writing for 24 hours, it is referred to as "active" in our technical documentation. We have a registry value that allows you to reduce this default period:
OracleRMANPluginReuseStorageTimeMinutes (HKEY_LOCAL_MACHINE\SOFTWARE\Veeam\Veeam Backup and Replication, reg_dword)
Could you please tell me how many Oracle servers and databases you have in your infrastructure? This information is important to prevent potential performance and scalability issues.
Thanks for this information. I was able to reduce to 12H and the copy from performance to capacity worked properly
I have usually up to 4 oracle databases all located on the same database server (between 1GB to 120 TB)
Rman perform the backup from 22H00 to 7H00 AM (maximum backup time frame) every day.
In terms of solution, my original idea was to replace the tape library (that provides me the 2d copy of my backup) by the S3 storage.
I firstly try to use the "backup copy job" but i found that this was not available for application backup plugin (oracle veeam plugin)
Then i saw the scale-out repository was supported as rman backup target. With the "copy" option it would have been the equivalent of my tape archival.
