understanding file sizes for backups and copy jobs

[rntguy]

So we have a single VM that we're backing up and copying that job offsite. I'm trying to understand how big the file really is and why.

VM1 - C Drive = 12GB in use out of 149GB; E drive 2.08TB in use out of 3.12TB. Was closer to 1.55TB in use a couple weeks ago.

VM1 - job says the VM is 2.1TB and the 'total size is 2.07 TB'. Weird but I'm guessing for the VM list to backup it's rounding to tenth place. And this is the 'used' space not the volume sizes.

I also assume based on the numbers below from 5/27 that this must have grown from 1.7 TB because of what you'll see below about the backup file size.

In Disk Properties for the job I can see this chain:
6/01 - VIB - Data Size: 1.65 GB; Backup Size: 725 MB
5/31 - VIB - Data Size: 3.49 GB; Backup Size: 1.76 GB
5/30 - VIB - Data Size: 3.33 GB; Backup Size: 1.98 GB
5/29 - VIB - Data Size: 1.06 GB; Backup Size: 409 MB
5/28 - VIB - Data Size: 1.03 GB; Backup Size: 390 MB
5/27 - FULL/VBK - Data Size 3.27 TB (volume size?) - Backup Size 1.7 TB

On 6/1 we started an Active Full seed copy job to the cloud. It just finished this morning, 6/8. It says it copied the restore point from 5/31, which I assume means it took the chain from 5/27 to 5/31 since 5/27 is where the FULL is that 5/31 feeds off of.

VM Size: 2.1 TB
DATA Processed: 2.1 TB
DATA Read: 2.1 TB
DATA Transferred 1.7 TB (1.2x)

However when I go to the Copy Job properties to see the chain and file sizes in the cloud I see:

vm-3845.287fD2017-06-01T104611.vbk
Data Size: 3.27 TB
Backup Size 2.06 TB

I thought this would show the size of the file at the backup repository. But it appears the BACKUP SIZE is showing the raw used TB amount on the volume, not the size of the backup.

So some questions:

How could the size of the file be 2.06 TB when it only transferred 1.7 TB?
Where do I see the actual size of the VBK file in Veeam software (either the backup job file or the copy job file post-backup)?
Does the active full on 5/31 really merge all the data from the full on 5/27 and all VIBs up to 5/31?

We have another similar file server that's bigger and we're trying to estimate how much space it will need for both the Full and 14 days of changes as well as the copy job full (just weekly fulls, no changes).

Thanks!

[dellock6]

Hi,
there are a couple of concepts you need to acquire before doing these calculations:
1- Veeam does backups at the block level, not the file level, so the size of the files in a volume may not be completely equal to the size of the backup.
2- we do compression and deduplication over a backup file, so the source size and the target size may be different.

This said, did you used by any chance the WAN accelerator for the backup job to the remote location? Because "transferred" means exactly data transferred from repo to repo, but it doesn't mean that some blocks were reused at destination by the destination cache of the wan-x for example...

for sizing, have you ever used the Restore Point Simulator?:
http://rps.dewin.me/

It may help you in these calculations.

Luca

[rntguy]

Hi.

No wan accelerator was used.

I get that block size will be bigger than file size depending on the files and the block size. It must be that it was able to do some of it's own compression and deduplication before sending it over. The source and target should be pretty close but the amount transferred over the wire could (and hopefully will) be smaller.

It's too bad the job properties says Data Size 3.27TB, backup Size 1.7TB since 3.27 TB is not really the amount of raw data, but the total size of the volumes. The amount of data in use is actually only 2.06 TB. It would be useful more so to see the amount of actual data you're protecting and then the backup file size. If you wanted to see the size of the volume i suppose that could be helpful but it's a bit misleading saying 'data size' when in reality it's volume size allocation not actual data size.

similarly in the backup job when it calculates the VM size it doesn't say 'used GB' just 'volume size'. Figuring out what that means was not simple from the documentation but checking a few against actual VMs you can see it's disk space in use.

[Vella]

Is it possible that 2 VM of about 300 GB, on disk occupy only 60/70 MB?

Is there a method to verify if a backup or a replica has been correctly performed?

Thanks, Bruno

[PTide]

Hi and welcome to our cozy community!

Answering your questions:

1. Yes, it is possible. In addition to Luca's explanation - if the VMs' data does not take all available space on the corresponding virtual disks and you've never used this unallocated disk space, then these blocks will be skipped.

2. Yes, there is such method - please check SureBackup.

Thanks

[Vella]

Thanks

But SureBackup is an option of Veeam B&R or another software?

[PTide]

Yes, it is an option of Veeam B&R, please check the link that I've provided in my previous post and let me know if you have any questions.

Thanks

[Vella]

Finally I have find where it is in the menu: first I need to define a Virtual Lab right?

Thanks for your useful help

[PTide]

That's right - you need to configure Virtual Lab first. Please check this guide for details on what Virtual Lab is, how it works, and how to configure it.

Thanks

[Vella]

Thank you very very much

But do you think it's realistic that a 300GB VM (about 200 GB used) became only 70 MB on disk?

My fear is that the backup in not done correctly

[PTide]

Oh, my bad, I actually have read it as 70 GB. Well, I have several assumptions regarding why that could have happened. Is there any chance that the VM has only RDM devices, or maybe the machines boot up from a network drive?

Thanks

[SBarrett847]

But do you think it's realistic that a 300GB VM (about 200 GB used) became only 70 MB on disk?

The only time I see something like that is with De-Duplication enabled in windows server, where Explorer does not show the true size of the data on disk, once dedupe is enabled..

[PTide]

Stephen,

I agree that deduplication can save a lot of space sometimes, however 2800x deduplication ratio (from 200 GB to 70 MB) must be some sort of a miracle

[foggy]

Bruno, how do you check the space that particular backup occupies on disk?

[Vella]

Directly in Windows folder

Can I post an image here?

[Vella]

Stephen,

I agree that deduplication can save a lot of space sometimes, however 2800x deduplication ratio (from 200 GB to 70 MB) must be some sort of a miracle

I'm agree with you Stephen

I'm a newbie with virtual environment but the compression seems a bit unreal

I would like to understand if it's realistic or there's something wrong

I don't want to discover it when I will need to restore something

[SBarrett847]

however 2800x deduplication ratio (from 200 GB to 70 MB) must be some sort of a miracle

no- it is that Windows does not report file sizes correctly with Dedupe enabled

EDIT:- Actually checking again, is the is the "Size On disk" that gets mis-reported. I have 100gb files, showing as Megabytes for Size on Disk.

The Bulk of the Dedup Blocks go to the System Volume Folder, and although these should be counted in the "Size on Disk", they are not. Also those blocks in the System Volume Folder are not Necessarily deduped either.

[patrickwilson412]

Does anyone know how to view the size of a vbk in the cloud through Veeam backup.

[PTide]

Please see the attachment.

Thanks