[ntp:questions] R: Re: debugging strange ntp in virtual environment

Riccardo Castellani ric.castellani at alice.it
Thu Sep 19 05:50:23 UTC 2013


>> This NTP server is a virtual machine (by P2V) but few months ago it was > 
working fine.
>Time control on a virtual machine is in general terrible. You 
should be
>having the underlying OS running ntp and having your virtual OS 
getting
>the time from that underlying OS.

So ntpd service, which gives time 
to clients, should be on physical machine ?
I ask myself if ntpd sets hardware 
clock too and in virtual environment if it changes hardware clock of ESX host.



>Remember that virtual servers have little control over the hardware, and
>get 
time at the whim of the underlying OS. Thus their system clocks need
>not be 
running at all regularly and they could have drift rates that are
>way outside 
of ntpd's range (500PPM) in which case ntpd will just give
>up.

how can I test 
my drift rate of virtual machine ?



More information about the questions mailing list