| 1 |
On Thu, Jul 30, 2020 at 2:52 AM Adam Carter <adamcarter3@×××××.com> wrote: |
| 2 |
>> |
| 3 |
>> > So should I run something like: date ; time <some command that runs at |
| 4 |
>> > 100%CPU for a minute> ; date ? |
| 5 |
>> |
| 6 |
>> No, date will pull from your RTC, which is usually kept up to date with an asynchronous |
| 7 |
>> counter. |
| 8 |
>> |
| 9 |
>> First check GNU top(1) and look in the %Cpu line for "st." That is % CPU time stolen. If it is |
| 10 |
>> nonzero then the guest time's accounting is probably working. It's not typical for the |
| 11 |
>> hypervisor to hide this information. It's really important for load balancing. |
| 12 |
> |
| 13 |
> |
| 14 |
> Thanks for that. I haven't seen any non-zero stolen time yet, however. |
| 15 |
> |
| 16 |
> FWIW vmstat also shows stolen time. |
| 17 |
|
| 18 |
Stolen time reporting through vmstat/top only works on xen and kvm |
| 19 |
hypervisors, it wasn't |
| 20 |
implemented for vmware. It actually looks like it was finally |
| 21 |
submitted for linux v5.7 |
| 22 |
(https://lore.kernel.org/lkml/20200331100353.GA37509@×××××.com/). |
| 23 |
|
| 24 |
If you want those numbers for older kernels, fetch this repository: |
| 25 |
https://github.com/dagwieers/vmguestlib |
| 26 |
and run vmguest-stats. You'll also need open-vm-tools installed. Have |
| 27 |
used this many times in the past and |
| 28 |
the numbers are good. |
Archives