Backup from a Windows File Server, with Microsoft deduplication enabled?

[Alexandre Guedes]

Hi Team.

I am looking for a Backup solution to backup something like ~10TB in total. One of the backuped servers will be a File Server (Windows) with ~5TB of data. The disk were this data is located has the Microsoft deduplication applied - https://docs.microsoft.com/pt-br/windows-server/storage/data-deduplication/overview - that help us to save something like 60% of storage.

My question is about the backup of this deduplicated disk. In my tests in the past with Bacula, I was able to run the full Backup without issues, however, the incremental backups never worked well. Seems that Bacula ended always runing a "full Backup", since the incremental backup size always was very similar with the full ones.

I was wondering if Bareos would handle well the Full+Incremental backups from a Windows deduplicated system. Tried to look for some documentation, but did not found any kind of Bareos documentation about.

Does you guys have this kind of environment (Windows deduplicated with Bareos + Full & Incremental jobs?). If yes, some advise about how to proceed?

Thank you

[Andreas Rogge]

Hi Alexandre,

Am 16.08.20 um 01:16 schrieb Alexandre Guedes:

> I was wondering if Bareos would handle well the Full+Incremental backups
> from a Windows deduplicated system. Tried to look for some
> documentation, but did not found any kind of Bareos documentation about.

it should just work. However, you will not see the deduplication in
Bareos, because it will take all the data so there will not be any
deduplication benefits there.

If your incrementals pick up too much data, it is probably best to
implement a small fileset (100 files or so) and try with that. This way
you can look at "list files jobid=XX" to see what files have actually
been backed up by the incremental and can provide a more meaningful
description of the problem, so we can find out what is going wrong.

Best Regards,
Andreas
--
Andreas Rogge andrea...@bareos.com
Bareos GmbH & Co. KG Phone: +49 221-630693-86
http://www.bareos.com

Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646
Komplementär: Bareos Verwaltungs-GmbH
Geschäftsführer: S. Dühr, M. Außendorf, J. Steffens, Philipp Storz

[Alexandre Guedes]

Hi Andreas,

Thank you for your answer. You are right. I ran some tests in my side and Bareos was able to backup the Windows files without issues. I've noted that when the file is sent to bareos, it already is hydrated. It does totally makes sense to me. So does not matter the state of the source (if it is duplicated or not), Bareos will always send the complete file.

Now, Just wondering how to apply the Deduplication in the Bareos side. I found this article https://docs.bareos.org/TasksAndConcepts/FileDeduplicationUsingBaseJobs.html and I am trying to understand this configuration in my scnerario.

In my side, the File server is unique, however, several files inside of this server are "duplicate". The JOB configuration that I am assuming to enable the dedu in Bareos side is:

Job {

      Name = StorageWin

      Level= Base

      ...

} 

Job {

      Name = StorageWinDeduplicated

      Base =  StorageWin

      Accurate = yes

      ...

}  

Does it makes sense? Some advise?

[Andreas Rogge]

Am 20.08.20 um 19:37 schrieb Alexandre Guedes:

>
> Does it makes sense? Some advise?

Right now there is no working deduplication in Bareos. AFAICT Basejobs
are not the way to go. They have some "interesting" (in the worst
meaning of the word) limitations when it comes to backup encryption,
copy & migrate jobs, and desaster recovery.
Sorry.