[ntp:questions] no drift-file on 2008 R2 vps and the time diff. is getting bigger and bigger?

William Unruh unruh at invalid.ca
Sun Sep 14 16:28:08 UTC 2014


On 2014-09-14, gooly <gooly at gmx.at> wrote:
>
> Sigh,
>
> Am 14.09.2014 11:00, schrieb William Unruh:
>>> Does this mean ntp will not even try to set the clock - it just gives up?
>> Yes. If the drift is greater than 500 it gives up.
>>
> It means I have to 'enrich' my own powershell-script (based on Chris' 
> script: 
> http://chrisjwarwick.wordpress.com/2012/09/16/getting-sntp-network-time-with-powershell-improved/) 
> for a continuous time setting.
>
No idea what that does

> So now I have some questions about the access of the ntp-server - may be 
> you can help me here.
>
> Btw I would have no problem to publish it - as soon as it is working
> with an enhanced time set more than just once every 30 Minutes or every 
> 15 min if a second higher limit is triggered.
>
> 1) I guess at least some ntp-server do have a limit to access. How often 
o
poll level 6 should be OK (Once a minute)

poll level 10 (once every 20 min) should certainly be fine


> npt is requesting the time, how does it deal with those limits?
>
> 2) My approach in my script is as follows:
o
This is completely silly. You ar trying for at verybest 1ms accuracy.
That is doaboe from any single ntpd source, ayou trying to badl do ntp
in your script is just dumb and irresponsible.

>I have a huge list of ntp-server around the world (well, mainly Germany 
> and Austria) and at first it requests the ntp-time from all, and sort 
> this server-list acc. to their delay of the request and then it takes 
> the 'fastest' for a check every 30 Minutes until the time difference 
> becomes bigger than a limit. Than again it checks all the nt-server and 
> choose the fastest which is in most cases another server.
>(comments, advices?)

Just use one.

>
> 3) Now I need to set 'continuously' the clock during the check interval:
> Let's take my vps: 7 sec delay per 15 min,
> that is 7.78 msec delay per 1 second.
> or 470 msc delay per 1 min.
> -> When would you set the clock using this drift?


What do you really need for accuracy. You cannot get accurac of rate,so
forget about that. certainly no more than once per min.


>
> Thanks in advance,
> Gooly



More information about the questions mailing list