@DeDa
Just following the reset procedures gives the same result.
But, I have got some new insights on the problem.
My setup uses 3PAR's to host al VM's.
Now one of the HP Best Practice is to deploy VM's with Eager Zeroed VMDKS's for the simple reason: have one place to manage your Thin Provisioning and well ... zero detect on the 3PAR rocks !
Apparently this messes up CBT on VMware.
So starting a VEPA backup with CBT enabled of a 50GB Eager Zeroed VMDK which only contains 10GB results in a 50GB backupdata.
So I went fiddling around with the VMDK format.
Making a VEPA backup of a 50GB Thin Provisioned VMDK which only contains 10GB results in a 10GB backup data.
Now convert the VMDK from Thin Provisioned to the preffered Thick Eager Zeroed format and follow the CBT reset mentioned by DeDA.
Result: 10GB of backup data.
Hmmmmmm... odd !
So resetting CBT while the original format is Eager Zeroed doesn't work, but after converting to thin and back to thick with CBT reset works !
On thing I notice is that VMware doesn't create its <vm-name>-ctk.vmdk file (in which it keeps track of the changed blocks) from the start of the VM after deployment.
However, at the start of the first backup with Data Protector it suddenly is there !
But, when executing the CBT reset procedure from VMware, the file is immidiatly recreated at the start of the VM.
So where do things go wrong ?
Converting all VM's from thick to thin and back to thick AND execting the CBT reset procedure is not a working option for me, since every VM has be powered of and on for a couple times for the CBT reset to work.
Any thoughts ?
Regards,
Martijn