Bottleneck varies considerably

[chimera]

Hi,

I don't understand how the bottleneck on a job can vary so considerably between 3 different VM's within the job.

Busy: Source 98% > Proxy 80% > Network 2% > Target 19%
Busy: Source 72% > Proxy 66% > Network 29% > Target 38%
Busy: Source 4% > Proxy 54% > Network 98% > Target 98%

All these VM's are backed up after hours (when they are doing nothing) one after the other, they all reside on the same LUN of a 10GbE iSCSI EqualLogic SAN (PS6010XV) and are all going to the same backup repository.

Can anyone explain the considerable variation? Is this normal?

Cheers

[Gostev]

Hi, different amount of changed blocks?