Full disk read every backup?

[TedStriker]

Hi,
I'm pretty new to Veeam and just trying out the free endpoint version on a 2012 server. I've set it to back up the entire computer to a local USB3 RDX drive and it seems to be working,however on every backup it is reading the entire contents of a disk and taking hours to back up just a few GB of changes.

Is this behaviour by design?

This is the entry in the job summary:

06/09/2016 00:33:36 :: 2TB (D:) (1.8 TB) 647.0 GB read at 115 MB/s [CBT]

There is about 650GB of data on that disk, very little of which would have changed since the previous night's backup. This portion of the backup takes 1.5 hrs

By contrast the System/C drive shows this:

06/09/2016 00:32:24 :: (C:) (117.0 GB) 7.0 GB read at 187 MB/s [CBT]

There is about 44GB of data on the C drive. This portion of the backup takes around 50 seconds.

Thanks,
Graham

[Dima P.]

Hi,

To clarify, what backup mode are you using? Thanks.

[TedStriker]

Hi, this is the job detail...

Backup job was modified successfully.

Source -------------------------------------------
Backup mode: [ entire computer]
The entire computer will be backed up

Destination --------------------------------------
Local folder: F:\VeeamBackup\

Schedule -----------------------------------------
Daily at [00:30]
If computer is shutdown at this time: [backup once powered on]
Once backup is taken, computer should: [keep running]
Backup no more often than every [2 Hours]

[Gostev]

I would guess you have a few of a very large files on disk D? What application?

[TedStriker]

Ok so some more information.

The D drive that is taking ages to back up is simple file storage. A mix of documents, music, video files and application installation images - total 81,000 files the largest of which is a clonezilla image of 103GB and the next one down is 10GB. Vast majority of files well under 1GB. No SQL databases etc. It contains virtually all static data, the only small changes would be in one directory synchronised by OneDrive. Now that does give me an idea - I'll try disabling the OneDrive client and see if that makes a difference.

One thing I have noticed though is that if I manually run the backup job shortly after a long run it only takes a matter of minutes on the second occasion.

[Dima P.]

Please open a case. I do remember that we have a fix for long merge issue and this might be related to your case. Do not forget to share the case ID in this thread. Thanks.

[adriand]

I see a similar thing on one server. Backup size looks small, says it uses CBT and reads the entire disks as shown in below image (it pretty much reads the entire existing data on the disks except drive C). Windows Agent version is 2.0.0.700. Was any solution found ? (not sure if it's same thing).

[adriand]

we had case 02308849 for this. response below:

"Veeam agent is entirely cluster unaware- https://www.veeam.com/kb2304 Presenting the drives as internal disks lets us leverage VSS against them, but underneath in the real world they're still cluster shared volumes. Veeam has no idea that any other nodes exist, but the data's constantly being accessed and replicated and manipulated between the nodes. This is expected behaviour for this set up. The blocks are changing as far as veeam can tell from looking at the windows master file table, because as far as windows is concerned, things are moving around and being changed."

long story short, this is actually a VM running Hyper-v 2012 R2 cluster (so on clustered volumes) and the VM itself is part of a DAG. We were forced to use the agent as currently we cannot snap the VM with the Veeam console (fails to take a snap of the VM at hypervisor level). But that it's a diff story.