Backing up Veeam - high Read value reported

[riaanb]

Hi

We do not have a large Veeam implementation (yet), I did however notice that at two of our sites two machines have very high read values reported for backups - in the first the provisioned size of the disks are 600GB, 596GB is read during each backup. In the second instance the provisioned disk size is 60GB and 56GB is reported as "read". on other VMs the read value is much lower - typically 1 or 2% of the actual disk space used. In the first case the "read" value represents 390% of the actual disk space used by the guest inside the VM, in the latter 175%. Thsi stays consistent for each backup of these machines.

Both VMs run Windows 2008, on ESXi 5.1 using thick, lazy zeroed virtual disks (as does the other virtual machine sin both environments) the only difference between these machines and the other machines is that Veeam B&R runs on these VMs.

Because pictures are w orth a thousand words I have attached one to this post.


Our understanding of the "read" value is that it represents the number of changes on the source VM before de-duplication or compression is done - is this correct?
Why is this read value so high for these machines?

[foggy]

Both VMs run Windows 2008, on ESXi 5.1 using thick, lazy zeroed virtual disks (as does the other virtual machine sin both environments) the only difference between these machines and the other machines is that Veeam B&R runs on these VMs.

If those VMs set up to backup themselves, then CBT is automatically disabled on them. This is a known limitation covered in the release notes. You can use another proxy server to back these VMs up.

[riaanb]

I did not see this in the most recent release notes and haven;t checked the older release notes.
Nevertheless, thank you - at least I know now.

[foggy]

Ah, you're right. There is a limitation on using hotadd while backing up virtual proxies, but no mention of CBT. Thanks for heads up.