Hi,

I have one S11 or Solaris 11 express system. All systems use ntp and 
connectes to the same ntp servers and using the same ntp.conf but the 
S11 always lags some 7 seconds from the other machines.

What has changed in ntp for S11 causing this, anyone know?

/michael

On 12/23/11 15:59, Michael Laajanen wrote:
> Hi,
>
> I have one S11 or Solaris 11 express system. All systems use ntp and
> connectes to the same ntp servers and using the same ntp.conf but the
> S11 always lags some 7 seconds from the other machines.
>
> What has changed in ntp for S11 causing this, anyone know?
>
> /michael

That sounds a bit weird. If the onboard clocks are hardware based (usually)
and you are getting regular updates, say daily, then the clocks should 
be within
a few milliseconds. Are you sure that the clocks really are getting 
updated, or
is the update interval not frequent enough ?. I use servers such as 
clock.isc.org,
time.nist.gov + others and can't remember having any problems. I guess 
it might
also be a subtle timezone problem...

Regards,

Chris

Hi,

ChrisQ skrev 2011-12-23 19:51:
> On 12/23/11 15:59, Michael Laajanen wrote:
>> Hi,
>>
>> I have one S11 or Solaris 11 express system. All systems use ntp and
>> connectes to the same ntp servers and using the same ntp.conf but the
>> S11 always lags some 7 seconds from the other machines.
>>
>> What has changed in ntp for S11 causing this, anyone know?
>>
>> /michael
>
> That sounds a bit weird. If the onboard clocks are hardware based (usually)
> and you are getting regular updates, say daily, then the clocks should
> be within
> a few milliseconds. Are you sure that the clocks really are getting
> updated, or
> is the update interval not frequent enough ?. I use servers such as
> clock.isc.org,
> time.nist.gov + others and can't remember having any problems. I guess
> it might
server ntp.nasa.gov
server time.nist.gov
server time-a.nist.gov
server time-b.nist.gov
> also be a subtle timezone problem...
>
> Regards,
>
> Chris
Hmm, what do you mean with a timezone problem?

/michael

On 12/23/11 19:05, Michael Laajanen wrote:

> Hmm, what do you mean with a timezone problem?
>
> /michael
>

Only insofar as they should be the same timezone,
not some equivalent in utc terms, but it's more
likely to be in the first two points above.

The ntpdc utility will display status info for each client and
should provide pointers on what's happening. For example:

$ ntpdc

ntpdc> sysinfo
system peer:          clock.isc.org
system peer mode:     client
leap indicator:       00
stratum:              2
precision:            -17
root distance:        0.18060 s
root dispersion:      0.02184 s
reference ID:         [204.152.184.72]
reference time:       d29f66d2.5e292000  Fri, Dec 23 2011 20:45:06.367
system flags:         monitor ntp stats calibrate
jitter:               -16.000000 s
stability:            7.383 ppm
broadcastdelay:       0.003906 s
authdelay:            0.000122 s
ntpdc> timerstats
time since reset:  2699783
alarms handled:    0
alarm overruns:    0
calls to transmit: 7377
ntpdc> loopinfo
offset:               -0.001413 s
frequency:            4.251 ppm
poll adjust:          30
watchdog timer:       1621 s
ntpdc>

and:

ntpdc> loopinfo
offset:               -0.001413 s
frequency:            4.251 ppm
poll adjust:          30
watchdog timer:       1621 s

man ntpdc for more info.

The clocks on both the solaris machines here are within a second of the 
radio msf
analog clock on the wall. Both machines running solaris 10...

Regards,

Chris

Hi,
ChrisQ skrev 2011-12-23 22:14:
> On 12/23/11 19:05, Michael Laajanen wrote:
>
>> Hmm, what do you mean with a timezone problem?
>>
>> /michael
>>
>
> Only insofar as they should be the same timezone,
> not some equivalent in utc terms, but it's more
> likely to be in the first two points above.
>
> The ntpdc utility will display status info for each client and
> should provide pointers on what's happening. For example:
>
> $ ntpdc
>
> ntpdc> sysinfo
> system peer: clock.isc.org
> system peer mode: client
> leap indicator: 00
> stratum: 2
> precision: -17
> root distance: 0.18060 s
> root dispersion: 0.02184 s
> reference ID: [204.152.184.72]
> reference time: d29f66d2.5e292000 Fri, Dec 23 2011 20:45:06.367
> system flags: monitor ntp stats calibrate
> jitter: -16.000000 s
> stability: 7.383 ppm
> broadcastdelay: 0.003906 s
> authdelay: 0.000122 s
> ntpdc> timerstats
> time since reset: 2699783
> alarms handled: 0
> alarm overruns: 0
> calls to transmit: 7377
> ntpdc> loopinfo
> offset: -0.001413 s
> frequency: 4.251 ppm
> poll adjust: 30
> watchdog timer: 1621 s
> ntpdc>
>
> and:
>
> ntpdc> loopinfo
> offset: -0.001413 s
> frequency: 4.251 ppm
> poll adjust: 30
> watchdog timer: 1621 s
>
> man ntpdc for more info.
>
> The clocks on both the solaris machines here are within a second of the
> radio msf
> analog clock on the wall. Both machines running solaris 10...
>
> Regards,
>
> Chris
>
Thanks for some hints, now works a blunder from me ofcourse :)

/michael