Sudden frequency offset jump - NTP

This is a discussion on Sudden frequency offset jump - NTP ; Hello, One of my systems had been running for ~6 weeks. During this time, the frequency offset computed by ntpd remained stable between -6.6 ppm and -6.2 ppm. On April 30, the time offset started climbing and reached 14 ms ...

+ Reply to Thread
Results 1 to 13 of 13

Thread: Sudden frequency offset jump

  1. Sudden frequency offset jump

    Hello,

    One of my systems had been running for ~6 weeks. During this time,
    the frequency offset computed by ntpd remained stable between
    -6.6 ppm and -6.2 ppm.

    On April 30, the time offset started climbing and reached 14 ms in
    approximately 90 minutes. ntpd started bumping the frequency offset,
    down to -9.4 ppm.

    Today, the frequency offset remains around -9.4 ppm.
    (And the time offset stays below 1 ms).

    # ntpq -crv
    assID=0 status=06f4 leap_none, sync_ntp, 15 events, event_peer/strat_chg,
    version="ntpd 4.2.4p0@1.1472 Fri Mar 16 10:45:43 UTC 2007 (1)",
    processor="i686", system="Linux/2.6.22.1-rt9", leap=00, stratum=2,
    precision=-20, rootdelay=45.975, rootdispersion=29.674, peer=33229,
    refid=134.226.81.3,
    reftime=cbc5889f.3df5b689 Fri, May 2 2008 12:33:35.242, poll=8,
    clock=cbc58c94.38f8bda6 Fri, May 2 2008 12:50:28.222, state=4,
    offset=0.099, frequency=-9.373, jitter=0.849, noise=0.699,
    stability=0.002

    One of the servers I use seems to have gone crazy:

    remote st t when poll reach delay offset jitter
    ================================================== ============
    212.27.63.123 5 u 235 256 377 27.929 759.246 1.475

    But I use several other servers, which all agree.

    Can someone offer an explanation for the sudden jump in the system
    clock's frequency offset?

    Regards.

  2. Re: Sudden frequency offset jump

    In article <481b11e9$0$14944$426a34cc@news.free.fr>,
    Noob writes:
    >Hello,
    >
    >One of my systems had been running for ~6 weeks. During this time,
    >the frequency offset computed by ntpd remained stable between
    >-6.6 ppm and -6.2 ppm.
    >
    >On April 30, the time offset started climbing and reached 14 ms in
    >approximately 90 minutes. ntpd started bumping the frequency offset,
    >down to -9.4 ppm.
    >
    >Today, the frequency offset remains around -9.4 ppm.
    >(And the time offset stays below 1 ms).


    Did the temperature change?

    My systems run about 1 PPM/C so it would take a pretty big
    shift to explain 3 PPM. Maybe somebody adjusted the
    air conditioner in the server room? Or installed a new
    server that's blowing a lot of hot air in your direction?

    --
    These are my opinions, not necessarily my employer's. I hate spam.


  3. Re: Sudden frequency offset jump

    Hal Murray wrote:

    > Noob wrote:
    >
    >> One of my systems had been running for ~6 weeks. During this time,
    >> the frequency offset computed by ntpd remained stable between
    >> -6.6 ppm and -6.2 ppm.
    >>
    >> On April 30, the time offset started climbing and reached 14 ms in
    >> approximately 90 minutes. ntpd started bumping the frequency offset,
    >> down to -9.4 ppm.
    >>
    >> Today, the frequency offset remains around -9.4 ppm.
    >> (And the time offset stays below 1 ms).

    >
    > Did the temperature change?


    The temperature in the server room is not constant at all, because the
    server room is cooled with air from outside. The temperature diff
    between night and day is on the order of 5-10C.

    Still, the system clock remained within -6.6 ppm and -6.2 ppm.
    (I do admit that I do not understand how that is possible.)

    > My systems run about 1 PPM/C so it would take a pretty big
    > shift to explain 3 PPM. Maybe somebody adjusted the
    > air conditioner in the server room? Or installed a new
    > server that's blowing a lot of hot air in your direction?


    The air conditioner is only run when the outside temp peaks over 25C.

    And someone did install a very hot server, but this was two weeks ago.

    To this day, ntpd considers the frequency offset has remained at
    approx. -10.1 ppm for over 36 hours.

    It looks like the crystal oscillator inside the system has been
    suddenly and permanently altered. Is that even possible?

    Regards.

  4. Re: Sudden frequency offset jump


    > It looks like the crystal oscillator inside the system has been
    > suddenly and permanently altered. Is that even possible?


    I believe that is quite possible as a result of aging.

  5. Re: Sudden frequency offset jump

    David Woolley wrote:
    >
    >> It looks like the crystal oscillator inside the system has been
    >> suddenly and permanently altered. Is that even possible?

    >
    > I believe that is quite possible as a result of aging.


    I believe that changes in crystal frequency due to aging are gradual
    and occur over periods measured in years.

    I would look, instead, at the other components of the oscillator!

  6. Re: Sudden frequency offset jump

    Richard B. Gilbert wrote:
    > David Woolley wrote:
    >>
    >>> It looks like the crystal oscillator inside the system has been
    >>> suddenly and permanently altered. Is that even possible?

    >>
    >> I believe that is quite possible as a result of aging.

    >
    > I believe that changes in crystal frequency due to aging are gradual
    > and occur over periods measured in years.
    >
    > I would look, instead, at the other components of the oscillator!


    Sudden jumps are not unknown, in addition to the more usual gradual aging.

    David



  7. Re: Sudden frequency offset jump


    Richard B. Gilbert wrote:
    > David Woolley wrote:
    >
    >>> It looks like the crystal oscillator inside the system has been
    >>> suddenly and permanently altered. Is that even possible?
    >>>

    >> I believe that is quite possible as a result of aging.
    >>

    >
    > I believe that changes in crystal frequency due to aging are gradual
    > and occur over periods measured in years.
    >
    > I would look, instead, at the other components of the oscillator!


    Like the power supply and associated power components.

  8. Re: Sudden frequency offset jump

    Andy Helten wrote:
    > Richard B. Gilbert wrote:
    >> David Woolley wrote:
    >>
    >>>> It looks like the crystal oscillator inside the system has been
    >>>> suddenly and permanently altered. Is that even possible?
    >>>>
    >>> I believe that is quite possible as a result of aging.
    >>>

    >> I believe that changes in crystal frequency due to aging are gradual
    >> and occur over periods measured in years.
    >>
    >> I would look, instead, at the other components of the oscillator!

    >
    > Like the power supply and associated power components.


    More like the components in the feedback loop!

    An oscillator is basically an amplifier with the output connected to the
    input; the squeal of a PA system with the volume set too high is the
    example familiar to almost everyone.


  9. Re: Sudden frequency offset jump


    Richard B. Gilbert wrote:
    > Andy Helten wrote:
    >
    >> Richard B. Gilbert wrote:
    >>
    >>> David Woolley wrote:
    >>>
    >>>
    >>>>> It looks like the crystal oscillator inside the system has been
    >>>>> suddenly and permanently altered. Is that even possible?
    >>>>>
    >>>>>
    >>>> I believe that is quite possible as a result of aging.
    >>>>
    >>>>
    >>> I believe that changes in crystal frequency due to aging are gradual
    >>> and occur over periods measured in years.
    >>>
    >>> I would look, instead, at the other components of the oscillator!
    >>>

    >> Like the power supply and associated power components.
    >>

    >
    > More like the components in the feedback loop!
    >
    > An oscillator is basically an amplifier with the output connected to the
    > input; the squeal of a PA system with the volume set too high is the
    > example familiar to almost everyone.
    >
    >


    Maybe I'm mistaken or maybe it was just a very cheap osciallator, but
    I'm fairly certain I've seen oscillator specs with voltage sensitivities
    shown in ppm/V. On such an oscillator, if a power supply changes the 5V
    or 3V output, it would change the frequency.

  10. Re: Sudden frequency offset jump


    >>> On April 30, the time offset started climbing and reached 14 ms in
    >>> approximately 90 minutes. ntpd started bumping the frequency offset,
    >>> down to -9.4 ppm.


    Is this the Linux TSC bug? Are you running a recent Linux kernel?
    Did you reboot the system at that time?

    --
    These are my opinions, not necessarily my employer's. I hate spam.


  11. Re: Sudden frequency offset jump

    Hal Murray wrote:
    >> Sudden jumps are not unknown, in addition to the more usual gradual
    >> aging.

    >
    > I don't have any numbers, but that seems like a very big jump.
    >
    > If it happened very often, I'd expect we would have seen similar
    > questions before.


    Hal,

    I didn't check the amount of the jump. I would not expect a jump of more
    than a few ppm (due to the crystal itself), and not more than about once a
    year.

    David



  12. Re: Sudden frequency offset jump


    >I didn't check the amount of the jump. I would not expect a jump of more
    >than a few ppm (due to the crystal itself), and not more than about once a
    >year.


    It was 3 or 4 ppm.

    That's big enough that I think it very unlikely, but not impossible.

    Even if it happened only once per year, somebody would probably
    have noticed by now. There are a lot of geeks on this list/newsgroup.

    Alternatively, there are probably other more likely explanations.

    --
    These are my opinions, not necessarily my employer's. I hate spam.


  13. Re: Sudden frequency offset jump

    Hal Murray wrote:

    > Noob wrote:
    >
    >> On April 30, the time offset started climbing and reached 14 ms in
    >> approximately 90 minutes. ntpd started bumping the frequency offset,
    >> down to -9.4 ppm.

    >
    > Is this the Linux TSC bug?


    No. I patched the kernel to make cpu_khz a compile-time constant.

    > Are you running a recent Linux kernel?


    2.6.22.1-rt9

    > Did you reboot the system at that time?


    No. The system has been running since 2008-03-21.

    Regards.

+ Reply to Thread