Linux Agent backing up more data than used?

[ASG]

Since I think i posted in the wrong sub I'll try my luck here:

Cross reference to this post: https://forums.veeam.com/physical-cloud ... 53043.html

Hi,

I'm using the Linux Agent (Server Lic) to backup one (later multiple, one is the test) Server over a WAN to our local B&R Repo. Agent was deployed and the Job is running managed from the B&R server.

Server is configured to backup indiviudal volumes (as there is a huge data volume I don't want to backup). However the Backup is larger than the source used space.

Since this is a WAN Backup i set High compression and optimization for WAN Target.

I configured /dev/sda1 (vFAT EFI partition /boot/efi with 500MB), /dev/sda2 (ext2 BOOT partition /boot with 485MB) and a LVM-Volume vg-rootfs (ext4 via /dev/mapper path with 200GB).

/ has about 7.2GB used space, the backupjob shows it has read 4GB at 29% - estimating about ~14GB or something like that in total. Process is running since this is a slow link for the initial Full-Backup

/boot and /boot/efi are about 160MB (together) used, the job log stated "Backed up sda 1000.5 MB at 1.4 MB/s" and the vbk was about 450MB.

This looks like it's backup up double the space? Is there anything I configured wrong? Do I have to switch to File-Level Backup? I thought since the agent is installed only used space is backed up?

Thanks!

Edit:
These are physical HPE ProLiant Servers, no virtual machines

I'm observing this on another Server, too (with entire computer selected):

/dev/sda1 (vFAT EFI partition /boot/efi with 500MB)
/dev/sda2 (ext2 BOOT partition /boot with 485MB)

Used space together is about 147MB, however while backing up the agent has transferred about 290MB. I don't understand this - the Agent does know about the Filesystem? Why does it backup double the space? The problem is - this Sever is sitting on a VERY slow link, like 800kbit/s upload SLOW

[PTide]

Hi,

Sorry for the long reply.

I configured /dev/sda1 (vFAT EFI partition /boot/efi with 500MB), /dev/sda2 (ext2 BOOT partition /boot with 485MB) and a LVM-Volume vg-rootfs (ext4 via /dev/mapper path with 200GB)

/ has about 7.2GB used space, the backupjob shows it has read 4GB at 29% - estimating about ~14GB or something like that in total. Process is running since this is a slow link for the initial Full-Backup

In other words, the resulting .vbk is around 14GB although the "/" has only 7.2 GB used out of 200GB, is that correct?

Thanks

[ASG]

Hi,

As far as I can tell yes - I deleted the Backup and decided to run another machine on a slower link (since this is the worst case link - it needs to work on that link) - I can't test it atm.

As i've edited I'm seeing something similar with a WAY worse connection. Almost the same sda1 and sda2 (since same servers installed from same media/image) - used space is ~147MB, but the Agent transfers ~415MB (which is a HUGE difference considering I'm doing the Backup at 500kb/s (not KiloByte - I'm talking about 500 kilobit a second ) - which results in a 480MB .vbk File on the Repo Server. That's almost 3x the used space on these volumes?!

Thanks!

Edit:


Edit2:
As stated in the other Thread - VAL can only skip XFS and ext4 unused blocks. So that's something I did miss - and that results in the larger backup for sda volumes. Thank!