Hey Guys,
I recently installed VE v5.1 onto an Intel NUC.. a fresh install, single host (non-clustered); and I've noticed some odd behavior on the CPU usage reporting for VM's, which were freshly created on this host.
The following screenshots are roughly the same time (within about two minutes of each other).
Here is my host:
Here is a Ubuntu Server 14.04 VM - Still in the 'pre-install' environment, sitting idle at the "specify hostname" prompt:
Here is a Ubuntu Server 16 VM - Still in the 'pre-install' environment, sitting idle at the "specify hostname" prompt:
And here is a Windows 10 VM, during the core-installation portion:
As you can see, the host, apollo2, is reporting a roughly constant 13% CPU usage for the first two VM's (both Ubuntu)... I have finished the installation for both VM's, this value will not fluctuate post-install.. and running 'top' reports nearly zero CPU usage in both Ubuntu environments.
What does cause the top two VM's to fluctuate on CPU usage (in Proxmox) for MISCVM1 (Linux) and MISCVM2 (Linux) is higher CPU usage on TESTWIN (Win10)... and it affects it in the inverse direction... as if CPU frequency is being ramped up, causing "overall" utilization to drop on the first two VM's, without the CPU demand (actually) dropping at all for those two VM's.
Can anyone explain this behavior, and if so, how can it be corrected? I don't know where Proxmox is getting it's 'baseline' from on these Linux VM's, but it is far from accurate. Since I've been writing this, the TESTWIN VM has dropped to an accurate ~2% CPU usage (waiting for Region selection in Setup).
Update:
Here are some additional screenshots:
Host running 'top':
One of the Ubuntu Linux VM's (MISCVM1) hardware and options:
Any help on this would be greatly appreciated!
I recently installed VE v5.1 onto an Intel NUC.. a fresh install, single host (non-clustered); and I've noticed some odd behavior on the CPU usage reporting for VM's, which were freshly created on this host.
The following screenshots are roughly the same time (within about two minutes of each other).
Here is my host:
Here is a Ubuntu Server 14.04 VM - Still in the 'pre-install' environment, sitting idle at the "specify hostname" prompt:
Here is a Ubuntu Server 16 VM - Still in the 'pre-install' environment, sitting idle at the "specify hostname" prompt:
And here is a Windows 10 VM, during the core-installation portion:
As you can see, the host, apollo2, is reporting a roughly constant 13% CPU usage for the first two VM's (both Ubuntu)... I have finished the installation for both VM's, this value will not fluctuate post-install.. and running 'top' reports nearly zero CPU usage in both Ubuntu environments.
What does cause the top two VM's to fluctuate on CPU usage (in Proxmox) for MISCVM1 (Linux) and MISCVM2 (Linux) is higher CPU usage on TESTWIN (Win10)... and it affects it in the inverse direction... as if CPU frequency is being ramped up, causing "overall" utilization to drop on the first two VM's, without the CPU demand (actually) dropping at all for those two VM's.
Can anyone explain this behavior, and if so, how can it be corrected? I don't know where Proxmox is getting it's 'baseline' from on these Linux VM's, but it is far from accurate. Since I've been writing this, the TESTWIN VM has dropped to an accurate ~2% CPU usage (waiting for Region selection in Setup).
Update:
Here are some additional screenshots:
Host running 'top':
One of the Ubuntu Linux VM's (MISCVM1) hardware and options:
Any help on this would be greatly appreciated!
Last edited: