Veeam One running on a virtual machine?

[sgatke]

Hi all
We are evaluating Veeam One for a small HyperV environment.
We would like to know if it is possible/advisable to run Veeam One on a virtual machine, on the same physical host that we are actually monitoring?

In our test, we found that e.g. CPU usage is very slowly updated in the Monitor.

Our test setup is like this:
HyperV on Server 2012.
3 virtual machines running:
VM1 - Server 2008 R2 (not really doing anything - low usage)
VM2 - Windows 7 Pro 64-bit (running CPU-intensive task, basically using 1 of 2 cores close to 100%)
VM3 - Windows 7 Pro 64-bit running Veeam One for HyperV

Infrastructure View - CPU shows that VM2 is using 44.00% CPU (which is about right).
However, if we stop the CPU-intensive task on the VM, the 44.00 are not updated, or at the very least, it takes many minutes before the numbers start to go down. When they go down, it is in small numbers like 44-38-34-etc. but with many minutes in between. Same can be seen on the Summary tab, with CPU Usage going from 1638 Mhz to 1588 Mhz after many minutes.
Is this behaviour on purpose or can we do something to get better response?

We expected a more direct reporting, like going from 44.00 to 8.00 or something like that.

This could of course happen because we run Veeam One virtualized on the same host as the one we monitor.
Just wondering...

[Vitaliy S.]

Hi Simon,

We would like to know if it is possible/advisable to run Veeam One on a virtual machine, on the same physical host that we are actually monitoring?

Both physical and virtual machines are supported as hosts for Veeam ONE server.

However, if we stop the CPU-intensive task on the VM, the 44.00 are not updated, or at the very least, it takes many minutes before the numbers start to go down. When they go down, it is in small numbers like 44-38-34-etc. but with many minutes in between. Same can be seen on the Summary tab, with CPU Usage going from 1638 Mhz to 1588 Mhz after many minutes.
Is this behaviour on purpose or can we do something to get better response?

When you stop these tasks, do you look at the performance graphs or on the summary page?

Thanks!

[sgatke]

Hi Vitaly

Thank you for taking the time to answer.

I am looking both places.
For Mhz usage, I am looking at the Summary tab.
For %-usage I am looking at the CPU tab (tried all views, but usually I start with "Past hour").

After having stopped the CPU-intensive task on VM2, I can see the local Windows Task Manager graph goes down to almost no CPU-usage.
It still takes several minutes until the CPU-graph in Veeam One recognizes the change.

I will try a reboot of all systems to night, and check again tomorrow if the issue is still present.

[Vitaliy S.]

Simon, I wouldn't recommend comparing in-Guest resource usage reported by Task Manager and Hyper-V monitoring API, instead you need to use Perf Monitor on the Hyper-V host to review VM performance.