Re: [ntp:questions] Question on how to Slew time in NTP (David Woolley)

2011-12-16 Thread David J Taylor
"Richard B. Gilbert"  wrote in message 
news:c5-dnee6pqwdtnftnz2dnuvz_o2dn...@giganews.com...

[]
Is this something relatively new?  ISTR that it still took about thirty 
minutes to get a reasonable facsimile of the correct time even when 
restarting with a good drift file.


I wish I could draw a graph here.  Since I can't, I'll try to describe 
it.  Start with an error of twenty or thirty seconds, mad dash to 
correct the error, followed by overshoot, and ringing.  Ringing decays 
slowly. If I were at the controls, I'd be tapping the brakes as we 
approach zero error!


I would try a recent NTP with configuration parameters appropriate to the 
environment you're using, with a valid drift file.  A 20s error should be 
corrected by a step, followed by fine adjustments.  I'm not talking 
microsecond offsets, though.


Cheers,
David 


___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions


Re: [ntp:questions] Large variation in NTP synchronization time

2011-12-16 Thread David J Taylor
"Joe Smithian"  wrote in message 
news:CA+pp-gKn-z-XmvbyqJ_ZOSz2YKAE+Ng_5iRRkQ0NDXrCL=8...@mail.gmail.com...

Hi all,

I've written a test script to test all the symmetric and autokey
authentication key types and algorithms. The scripts  waits for a delay
time after restarting ntpd and then runs the "ntpq -c as" command to see
if  '*' has appeared beside the server name.The problem is that the
synchronization time varies from about 20 seconds to more than 15 
minutes,

even when using -burst and -iburst options for a private ntp server!
What are the minimum, average and maximum synchronization times?
Is there a faster way to verify synchronization?

Thanks

Joe


Joe,

I would verify sync with "ntpq -p".

On my systems, I would expect to see a "*" almost immediately after 
starting NTP, and if I didn't see it within (say) 20 seconds I would 
investigate.  However, I'm not using any authentication.  I hope that 
helps.


Cheers,
David 


___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions


Re: [ntp:questions] Configure FreeBSD or Linux to use stepping clock?

2011-12-16 Thread Miroslav Lichvar
On Thu, Dec 15, 2011 at 07:50:08PM +, Dave Hart wrote:
> Dr. Mills raised the possibility privately that either FreeBSD or
> Linux might be reconfigured to use a more primitive clock that steps
> once per millisecond or less.  If possible and I am able to accomplish
> it, my testing of these bug 2037 fuzzing changes would be greatly
> assisted.

On Linux, you could set a different kernel clocksource. Perhaps
to jiffies or pit, if available.

Check these files:
/sys/devices/system/clocksource/clocksource0/current_clocksource
/sys/devices/system/clocksource/clocksource0/available_clocksource

Why not degrade the resolution of the clock directly in ntp sources?
In get_systime():
GET_SYSTIME_AS_TIMESPEC(&ts);
ts.tv_nsec /= 100;
ts.tv_nsec *= 100;

-- 
Miroslav Lichvar
___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions


Re: [ntp:questions] Configure FreeBSD or Linux to use stepping clock?

2011-12-16 Thread David Woolley

E-Mail Sent to this address will be added to the BlackLists wrote:

Dave Hart wrote:

Linux might be reconfigured to use a more primitive clock that steps
once per millisecond or less.


Linux Kernel 2.4 used 100 Hz and 2.6 increased it to 1000 Hz?


They still read the time to the full CTC resolution, so had precisions 
corresponding to something more like a microsecond.


___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions


Re: [ntp:questions] How to verify Autokey Identity Schemes?

2011-12-16 Thread Joe Smithian
Hi Steve,

I am posting my questions again in text format. I hoe you can read it this time.

Thank you for your comments. I tried ntpq -c "rv assID flags" command,
it shows the
Identity Scheme that the server supports regardless of what identity
scheme has been installed on the client.
Here are the result of my experiments:

Server Identity scheme    | ntpq -c "rv assID flags"
-|--
IFF    |  0x417f21
GQ    |  0x417f41
IFF and GQ    |  0x417f61

"rv assID flags" returns the same value whether I install IFF
parameters, or GQ parameters or none on the client. So my question
again is that how can I verify that IFF or GQ schemes are actually
working?

Association flag shows auth is 'ok' whether I install an Identity
Scheme on the client or not, so it's not an indication that IFF or GQ
is actually being used.

BTW, I found two problems in this
document:http://support.ntp.org/bin/view/Support/ConfiguringAutokey

In sections 6.7.2.5 and 6.7.3.6:
    ntp-keygen -T -q `awk '/crypto pw/ { print $3 }'  wrote:
>
> Hi Steve,
>
> Thank you for your comments. I tried ntpq -c "rv assID flags" command, it 
> shows the
> Identity Scheme that the server supports regardless of what identity scheme 
> has been installed on the client.
> Here are the result of my experiments:
>
> Server Identity scheme    | ntpq -c "rv assID flags"
> -|--
> IFF    |  0x417f21
> GQ    |  0x417f41
> IFF and GQ    |  0x417f61
>
> "rv assID flags" returns the same value whether I install IFF parameters, or 
> GQ parameters or none on the client. So my question again is that how can I 
> verify that IFF or GQ schemes are actually working?
>
> Association flag shows auth is 'ok' whether I install an Identity Scheme on 
> the client or not, so it's not an indication that IFF or GQ is actually being 
> used.
>
> BTW, I found two problems in this 
> document:http://support.ntp.org/bin/view/Support/ConfiguringAutokey
>
> In sections 6.7.2.5 and 6.7.3.6:
>     ntp-keygen -T -q `awk '/crypto pw/ { print $3 }' 
>   '-q' option for updatuing keys doesn't work, '-p'works; is this a typo 
> in the document?
>
> [root@myserver]# ntp-keygen -T -q `awk '/crypto pw/ { print $3 }' 
>  Using OpenSSL version 90802f
> Using host myserver group myserver
> Corrupt file ntpkey_host_myserver or wrong key myserver
> error:06065064:digital envelope routines:EVP_DecryptFinal_ex:bad decrypt
>
> Regards
>
> Joe
>
>
>
> On Tue, Dec 13, 2011 at 10:55 AM, Steve Kostecke  wrote:
>>
>> On 2011-12-12, Joe Smithian  wrote:
>>
>> > I have configured my NTP server and client to use Autokey with IFF
>> > Identity scheme and it's working, client synchronizes to my servers.
>> > It synchronizes with and without copying the IFF parameter to the
>> > client. So I'm wondering if IFF identity scheme is actually being
>> > used; How can I verify that?
>>
>> By checking the association flags.
>>
>> Please see
>> http://support.ntp.org/bin/view/Support/ConfiguringAutokey#Section_6.7.4.
>>
>> --
>> Steve Kostecke 
>> NTP Public Services Project - http://support.ntp.org/
>>
>> ___
>> questions mailing list
>> questions@lists.ntp.org
>> http://lists.ntp.org/listinfo/questions
>
>
___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions


Re: [ntp:questions] How to verify Autokey Identity Schemes?

2011-12-16 Thread Steve Kostecke
On 2011-12-16, Joe Smithian  wrote:

> I tried ntpq -c "rv assID flags" command, it shows the Identity Scheme
> that the server supports regardless of what identity scheme has been
> installed on the client.

NTP Authentication is used to authenticate the server to the client. So
you need to look at the flags on the client.

-- 
Steve Kostecke 
NTP Public Services Project - http://support.ntp.org/

___
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions