oversized vzdump backup

[jans16]

I use the backup function on the web-interface.
But in some of the vm's the vzdump is very oversized:
When I look with the df command on the vm I get the information that 1.8 GB of the 20 GB vhd are used. But the backup will need 11.15 GB on hd (said the backup status-mail).
On an similarly vm is 1.2 GB in use, the vzdump only need 1.4 GB, thats ok, I think.

It takes a lot of time to copy the oversized backup to an external usb-hd (more than 1 hour) in contrast to the "normal" backup-size which need only 10 minutes.

I think this is a problem with freespace on the vhd (on an LVM-group), that was not free in the past. So the virtualized system has marked the space as free but the host don't know it. Knows anyone a possibility to make freespace on vhd realy free, so that this is not backed up? Or is there any other solution?

Thanks a lot.

 

[lp86]

Sounds like this is a KVM machine. You need to zero and shrink the disk image file. As you add data the disk image expands, but even when you delete it the data is still there.

 

[jans16]

Yes, it's a KVM machine.
And it is not an Imagefile, but an LVM Logical Volume.
How can I zero (and shrink) the virtual disk?

 

[Gigulon]

Hey Jans16,

did you get any answer for your Problem ?
I have the same Problem. Can anybody help ?

In my KVM runs an windows 2008 Server R2 and i deleted about 100Gig of data but the backup still remains on the old size.

Thanks
Peter

 

[tom]

Read also here: http://forum.proxmox.com/threads/3988-how-to-shrink?p=23766#post23766

 

[jans16]

My problem is still unresolved.
On my vm runing a Ubuntu Server. I found a way to zero the free disk space on the vm:

dd if=/dev/zero of=zero.small.file bs=1024 count=102400
dd if=/dev/zero of=zero.file bs=1024
sync ; sleep 60 ; sync
rm zero.small.file
rm zero.file

But now, the backup have the full size of the Logical Volume, that is used as vm-disk!

@tom: this link is only for using image-files, but not for a LV of LVM.