thoughts about backup with Oracle Archive S3 Storage after migration from Veeam

Mar 15, 2026
5
0
1
Hi,

currently I'm running a VMware ESXi (single server) with Veeam VM, backup with forward incremental to a Qnap. The Backupdatastore on Qnap get synced daily to an Oracle S3 Storage.
Once a month Veeam is running a full backup on each job, afterwards the forward incremental continues.
On Oracle two rules exist:
1. after 24h of last write the files get migrated by Oracle from S3 to Archive Storage
2. the S3 Archive has a retension of 91 days (to have minimum retension time)

All restores was till now possible from local Qnap.

In case of disaster, after having organized some hardware.
1. a restore from Oracle Archive to Oracle S3 must get requested (can take up to 24h)
2. last full-backup (max age 30days) + all following incr. must get copied to local system (+ one metadata file per VM)
3. Restore of this files with Veeam can happen to new hardware


When migrating this to a PBS (running native on a ugreen NAS) I try to come to a similar approach, but currently I'm unable to find a working solution, cause all
* PBS don't know about the inaccessability for chunk files on Archive storage
* I don't know, which chunk files are required (in case of required restore request from archive to normal)
* Oracle retension has no information about the requirement of a file based on deduplication from PBS (would be required for autodelete after expiry of retension time)
* when having everything on Objectstorage
- data are not necessary protected from deletetion.
- costs are 10x higher (might be superseeded with a better deduplication, which was explicitly denied for per VM backup files on Veeam)

any ideas about this?

br
Tom


https://www.oracle.com/de/cloud/storage/pricing/
 
hm, currently I'm experimenting with the S3 Techpreview and see some pros and cons.
1. creating an S3 endpoint + afterwards an datastore is not intuitive in the beginning, but I guess there is some fusefs underneath, when seen, remembered.
2. have the possibility to have only encrypted and verified jobs uploaded is really good.

3. where I'm a bit struggling, I had a retension policy set to prevent overwriting a file in the S3 storage, this prevented me in creating the PBS datastore.

This I see a bit more critical - even this is directly related the Storage Archive - cause a compromised PBS, can the delete the offsite backup as well.

From that point of view I'm still looking for a 3-2-1 backup, but currently it looks it's only possible with a second PBS, which does a pull job ;(

br
Tom