File structure, S3 Backup copy vs Local Repo - totally different?

[mc625569]

Hello,

We are testing SOBR with s3/immutable for the first time. I've created a SOBR, and set it to immediately copy to the capacity tier with a small test vm.

The job is showing complete with a single 15GB vm. When I check the performance tier repository I see a single*.vbk file showing roughly 15gb. When I browse the s3 repo via the amazon interface I see the same 15GB used, but as thousands of *.blk files and somewhat different folder structure.

two questions:
1. Does the capacity tier repo always have a different folder structure vs. the the performance tier, namely no *vbk file ?
2. How can we test a restore from s3 to make sure it is working properly? When we choose a restore job we do not have a choice between which of the 2 parts of the SOBR to choose.

[HannesK]

Hello,

1. Yes, it's completely different. Object storage is different than NAS or block storage simply because there's no file system to start with. Because of that, there are no Veeam backup files. Please see the user guide: https://helpcenter.veeam.com/docs/backu ... ml?ver=100

2. I usually put the performance tier in "maintenance mode": https://helpcenter.veeam.com/docs/backu ... ml?ver=100

As an alternative, you could also use a second VBR server and import the backups.

Best regards,
Hannes

[pirx]

I just found this old thread.

#2 is something that should be improved in my opinion. One might want to restore from S3 for a reason, putting the local extent in maintenance is not always an option. In our case we want to test disaster recovery from local storage and from S3 for a large number of VM's. But we are unable to choose to restore from S3 without any workaround. It'd really make sense to let the customer choose from which tier he might want to restore.

[veremin]

2. You can select backups in object storage node and initiate restore from there. However, if you're restoring to production environment (not public cloud), backup server will still try to re-use data on performance tier to optimize restore speed and cost.

To simulate lost of Performance Tier you will need to put performance extents into maintenance mode, as Hannes suggested. Then, backup server will use only blocks stored in object storage.

Anyway, thank you for the feedback, your voice has been heard.