Skip to main content

If you set your S3 repos gateway server to direct mode, I would have expected the entity getting in touch with backup data to do the offload.

In direct to object jobs it works just like that. Proxy will do the upload with a primary job. Repo will do the upload with a backup copy job. 

But with SOBR offloading to capacity tier it seems to be the VBR server doing the offload for all S3 buckets in direct mode.

Can someone confirm this observation?

Would be another reason to always have GW servers specified for each and every S3 repo.

https://helpcenter.veeam.com/docs/backup/vsphere/gateway_server.html?zoom_highlight=Direct+mode&ver=120
 

The expected behaviour you described is what I had witnessed whenever I’ve used it, can you check what your mount server is please?

 

From the documentation:

  • Object storage repository with the direct connection mode. The mount server associated with the source backup repository is used as the gateway server. If the mount server is not accessible, Veeam Backup & Replication fails over to the backup server.


Thanks, @MicoolPaul. You nailed it - for S3 as primary.

I mixed it up with DAS as primary. Then you have the DAS repos themselves become the gateway for the upload - as I would have expected.

It’s getting complicated with these different repo types. So, better always define your GW. 😊


Comment