Immutable Repo - Capacity / Free Size difference

[david.brunner]

Hi forum members!

I installed a new Veeam v12 B&R server and used the old one to install as Ubuntu Server and use it (with XFS) as an immutable repo. I used https://helpcenter.veeam.com/docs/backu ... ml?ver=120 as guide (but not finished the hardening yet)
for now I want to do some tests, if it works and to estimate, what I copy on it and how long ...

What I do not understand is the capacity information.

Backup Infrastructure -- Backup Repositories
In the view plane, I see
Name | Type | Host | Path | Capacity | Free | Used Space
Imm... | Hardened | 10....| /mnt/... | 16,4 TB | 10,5 TB | 11,8 TB

can someone tell me, how that sums up?
Free + Used Space is over the overall capacity.

sorry, if this a really simple, redundant question
Veeam does a great job in documenting stuff. But the more is written, the less you might find...

[Mildur]

Hello David

I moved the topic from the object storage forum. Hardened repositories are not object storages.
I assume you are using XFS as the filesystem. In that case, our jobs can use FastClone to create synthetic backup files.

When you create a synthentic full backup, the backup job will use data blocks from existing backup files and changed/new blocks from the current job run to synthesize a new full backup file.
- Without FastClone, we have to copy all data blocks and create a new file.
- With FastClone, we don't need to copy data blocks. We just create a new file and reference to the data blocks already on disk. This gives you almost space less full backups. Only changed/new data will be written to the repository. You will save a lot of storage and time with synthetic full backups.

To answer your question, used Space you see in the backup console will show you the from the file system reported size of all backup files. Reported file size is always the entire file size. The filesystem is not aware about the space savings/referenced data blocks.

Userguide about FastClone:
https://helpcenter.veeam.com/docs/backu ... ml?ver=120
Blog post from @jorgedlcruz :
https://jorgedelacruz.uk/2020/03/19/vee ... -in-veeam/
Youtube video explaining FastClone. For ReFS, but it's the exact same behavior for XFS:
https://www.youtube.com/watch?v=NndMBCDPBDY

Best,
Fabian

[david.brunner]

Hi Fabian,

as a hopelessly overworked and sometimes clueless technician/user I thank you...
so it is "deduplicatioon" and I understand
I might have understood if I found the information, but like I wrote, it is a "info overload" and do not know where/how to start

yes, I use XFS and then your information is the answer for that

THANK YOU! For the info, for your time.

best regards from Austria

David

[tyler.jurgens]

The data isn't de-duplicated. In nontechnical terms, it simply uses XFS's ability for identical blocks to be pointers to a single piece of data, rather than duplicating the data every time. The data is not de-duplicated, its never duplicated to begin with.