Veeam CBT Processing Full Drive

[JeffN825]

We are evaluating Veeam Backup and Replication. I set up Veeam to back up our VMWare server about two weeks ago.

I've seen some threads about CBT not working because of snapshots and that causing backup jobs to process the full hard drive...but in my case it seems like CBT actually IS working.

The problem started just under a week ago - for full backups of VMs our VMWare ESX 5.1U1 server, Veeam (located on a different physical server - eg using Network Backup) is processing and reading the entire drive capacity of the source VMs. This applies to all VMs on the source server.

The backup log looks something like this (for each VM):
Hard Disk 1 (100.0 GB) 99.3GB read at 52MB/s [CBT]

A week ago, a similar full backup read (for the same source VM):
Hard Disk 1 (100G.0 GB) 15.1 GB read at 39MB/s [CBT]

Needless to say, this is a huge problem, as full backups now take 12+ hours. How can I go about troubleshooting this?

Thank you.

[kjc3303]

I had a similar issue on a particular VM may have even been a single vmdk? I found that resetting the CBT solved the issue. Seems a bit strange that you are having this issue on all VM's so may be a different issue

Kev

[foggy]

I've seen some threads about CBT not working because of snapshots and that causing backup jobs to process the full hard drive...

That's true, in cases where there is at least one existing snapshot, CBT cannot be enabled on the VM. However, if CBT has been enabled prior to creating a snapshot, Veeam B&R will be able to successfully utilize CBT.

but in my case it seems like CBT actually IS working.

It is enabled for this VM, however returns a huge amount of changed blocks for some reason. Is there anything that could cause huge amount of changes inside your VMs (defragmentation or antivirus scan for example)?

That said, in some rare cases CBT data can get corrupt and trying to reset CBT could be the first step in addressing this issue.

Btw, what about incremental runs? Do you observe such behavior on full job runs only? Then the cause could be in dirty virtual disk data blocks belonging to the deleted files.

[JeffN825]

Resetting CBT fixed it. Thanks.