Backup performance question

[jleclerc]

I'm testing out a new Veeam B&R environment and I'm confused with big performance differences between 2 backup jobs.

The two job are similar, in fact the second one is a clone of the first one. Same source infrastructure, same proxy, same networking path, and same target. Both have AAP and VMware Quiesce disabled and use the same compression/deduplication settings. Both of the values below are based on first backup run on each job (Active Full).

The first job gets 139MB/s with the following load: Source 80% > Proxy 86% > Network 75% > Target 32%
The second job gets 914MB/s with the following load: Source 86% > Proxy 83% > Network 72% > Target 21%

The only difference is VMs being backed up. Job 1 is backing up 5 Windows Server VMs while Job 2 is backing up 3 Linux VMs

Why would I be getting 6.5x performance boost on the 2nd job?

[ronnmartin61]

Are they processing comparable amounts of data?

[jleclerc]

Job 1 about 590GB of data, Job 2 about 480GB, so similar enough.

[ronnmartin61]

Best bet is to open a case with support

[jleclerc]

Sounds good, thank you.

[Origin 2000]

You should compare the amount of "Read" data within the job stats and needed time. i assume thin? vDisk and your Windows VMs may have a much higher fill ratio as the Linux ones.

For example:
VM 1: 10TB and 9TB data inside
VM 2: 20TB and 1TB data inside