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

unruh unruh at invalid.ca
Fri Sep 20 03:54:25 UTC 2013


On 2013-09-19, Rob <nomail at example.com> wrote:
> unruh <unruh at invalid.ca> wrote:
>> On 2013-09-19, Horvath Bob-BHORVAT1 <Bob.Horvath at motorolasolutions.com> wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: questions-bounces+bob.horvath=motorolasolutions.com at lists.ntp.org
>>>> [mailto:questions-bounces+bob.horvath=motorolasolutions.com at lists.ntp.org] On
>>>> Behalf Of Rob
>>>> Sent: Thursday, September 19, 2013 11:20 AM
>>>> To: questions at lists.ntp.org
>>>> Subject: Re: [ntp:questions] R: Re: debugging strange ntp in virtual
>>>> environment
>>>> 
>>>> Horvath Bob-BHORVAT1 <Bob.Horvath at motorolasolutions.com> wrote:
>>>> >
>>>> 
>>>> Why?
>>>> What we have is a similar situation, we had a physical machine providing ntp
>>>> service to the network and we went to virtualisation.
>>
>> Why why why would you want to do that? It is simply nuts. 
>
> Because the boss wants to cut down on physical machines, as we are
> migrating to external hosting.

Oh well. You are having trouble. You wonder where it is coming from.
Suggestions are made and you reject them. Fine.

>
> Unruh, what we see time and time again: you as a hobbyist on your own
> little PC at home have a completely different view on the world than the
> corporate sysadmin has.

OK, except I do admin a group of about 8 machines at my work.

>
> You are always recommending things that just aren't practical.
> Just as I can't get a Sure GPS board and install its antenna outside
> the building with a long wire to the computer room, I also cannot decide
> if we will virtualize or not.

And I know what is practical how? You ask for help. Suggestions are made
and it is up to you whether or not to impliment them. But, sometimes
refusing them comes with a cost. So you will have to decide if the cost
is worth it or not. 


>
> We virtualized about 30 servers into 3 boxes at one location, and 4
> servers into another box at another location.  The footprint went from
> a room full of equipment to 6HE in a rack.
> Having a few hundred us less NTP accuracy is a small price to pay for
> that.

That is or course your decision. And whether it is a few hundred us or a
few hundred ms, or sec will depend on how busy your virtual machines
are. If they get busy enough, ntpd will not be able to handle the wild
fluctuations in drift rate. 

Why do you not have your ntp servers running on the underlying OS,
instead of on your virtual machines?
 
>
> If only you would understand such things...

I do. But if only you would understand that the information you have
given above was NOT in your original request. 
>
>>>> The machine still providers good time.  It is still synchronized to time
>>>> sources on the network as it was before, and it still provides time to other
>>>> servers and clients on the local network.  Servers are other virtual servers
>>>> on the same ESX hosts.  Loading of the origin time servers is the same.
>>
>> It seems to now, and then that machine becomes loaded, with one of the
>> other virtual machine taking up loads of time, and suddenly your time
>> has gone to hell. This is how you want a server to behave?
>
> You have zero knowledge of how ESX behaves.
> Probably you base your opinion on a VMware workstation installation you
> once did on your hobby PC.
> Please don't make statements about stuff you have not the slightest
> knowledge of.
>
>> Sure it works since you test it on a lightly loaded machine where each
>> OS gets approximately the same amount of time. 
>
> And also when it is loaded.  That is why you have an 8-core machine
> underneath it, and a virtualization platform that allows you to allocate
> resources.



More information about the questions mailing list