timekeeping on Soekris net4801 w/ ntpd. 3.8

classic Classic list List threaded Threaded
53 messages Options
123
Reply | Threaded
Open this post in threaded view
|

timekeeping on Soekris net4801 w/ ntpd. 3.8

J Moore
I just installed 3.8 on a Soekris net4801 that's been laying around for
a while (unused, unpowered). I noticed after install that time was off
by like 5 months, so I set it to within a few minutes of current
time/date from the wall clock.

I've been checking the logs, and this is what I'm seeing... this has
been going on for about 8 hours now. Why is ntpd having to make 60+
second adjustments every 3-5 minutes? It would appear the clock on the
Soekris is really BFU.

I researched this a bit, and found out there were (are) some issues with
timekeeping on the net4801's processor, but the other platforms (Linux,
FreeBSD) seemed to have solved the problem by using a different clock. I
saw nothing that would suggest any current problems with OpenBSD.

WTF, O?

Thanks,
Jay



Nov 14 06:30:10 opie ntpd[4133]: adjusting local clock by -91.931803s
Nov 14 06:34:22 opie ntpd[4133]: adjusting local clock by -90.983786s
Nov 14 06:37:42 opie ntpd[4133]: adjusting local clock by -90.136183s
Nov 14 06:38:01 opie ntpd[14058]: peer 213.61.224.44 now valid
Nov 14 06:41:25 opie ntpd[4133]: adjusting local clock by -88.969563s
Nov 14 06:44:33 opie ntpd[4133]: adjusting local clock by -88.039023s
Nov 14 06:45:45 opie ntpd[14058]: peer 65.102.104.139 now valid
Nov 14 06:48:48 opie ntpd[4133]: adjusting local clock by -87.275355s
Nov 14 06:52:15 opie ntpd[4133]: adjusting local clock by -86.527847s
Nov 14 06:55:43 opie ntpd[4133]: adjusting local clock by -85.775024s
Nov 14 06:57:38 opie ntpd[14058]: peer 65.102.104.139 now invalid
Nov 14 06:58:19 opie ntpd[4133]: adjusting local clock by -85.221285s
Nov 14 07:01:38 opie ntpd[4133]: adjusting local clock by -84.338437s
Nov 14 07:04:48 opie ntpd[4133]: adjusting local clock by -83.437080s
Nov 14 07:08:18 opie ntpd[14058]: peer 65.102.104.139 now valid
Nov 14 07:08:40 opie ntpd[14058]: peer 65.102.104.139 now invalid
Nov 14 07:08:44 opie ntpd[4133]: adjusting local clock by -82.500680s
Nov 14 07:12:15 opie ntpd[4133]: adjusting local clock by -81.562691s
Nov 14 07:15:44 opie ntpd[4133]: adjusting local clock by -80.871930s
Nov 14 07:19:24 opie ntpd[4133]: adjusting local clock by -80.103985s
Nov 14 07:19:55 opie ntpd[14058]: peer 65.102.104.139 now valid
Nov 14 07:22:17 opie ntpd[4133]: adjusting local clock by -79.112030s
Nov 14 07:26:17 opie ntpd[4133]: adjusting local clock by -78.507867s
Nov 14 07:27:26 opie ntpd[14058]: peer 65.102.104.139 now invalid
Nov 14 07:30:31 opie ntpd[4133]: adjusting local clock by -77.137970s
Nov 14 07:34:31 opie ntpd[4133]: adjusting local clock by -76.275602s
Nov 14 07:38:22 opie ntpd[14058]: peer 65.102.104.139 now valid
Nov 14 07:38:44 opie ntpd[4133]: adjusting local clock by -75.556709s
Nov 14 07:41:22 opie ntpd[4133]: adjusting local clock by -74.659299s
Nov 14 07:41:38 opie ntpd[14058]: peer 65.102.104.139 now invalid
Nov 14 07:44:40 opie ntpd[4133]: adjusting local clock by -73.778046s
Nov 14 07:47:49 opie ntpd[4133]: adjusting local clock by -73.123884s
Nov 14 07:52:09 opie ntpd[4133]: adjusting local clock by -72.233839s
Nov 14 07:56:05 opie ntpd[4133]: adjusting local clock by -71.240807s
Nov 14 07:59:08 opie ntpd[4133]: adjusting local clock by -70.277551s
Nov 14 08:02:27 opie ntpd[4133]: adjusting local clock by -69.496544s
Nov 14 08:06:30 opie ntpd[4133]: adjusting local clock by -68.464538s
Nov 14 08:10:36 opie ntpd[4133]: adjusting local clock by -67.807755s
Nov 14 08:14:14 opie ntpd[4133]: adjusting local clock by -66.636277s
Nov 14 08:14:23 opie ntpd[14058]: peer 65.102.104.139 now valid
Nov 14 08:17:56 opie ntpd[4133]: adjusting local clock by -66.035356s
Nov 14 08:20:36 opie ntpd[4133]: adjusting local clock by -65.174887s
Nov 14 08:24:20 opie ntpd[4133]: adjusting local clock by -64.294286s
Nov 14 08:27:59 opie ntpd[4133]: adjusting local clock by -63.612736s

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Alexander Hall
J Moore wrote:
> I just installed 3.8 on a Soekris net4801 that's been laying around for
> a while (unused, unpowered). I noticed after install that time was off
> by like 5 months, so I set it to within a few minutes of current
> time/date from the wall clock.
>
> I've been checking the logs, and this is what I'm seeing... this has
> been going on for about 8 hours now. Why is ntpd having to make 60+
> second adjustments every 3-5 minutes? It would appear the clock on the
> Soekris is really BFU.

> Nov 14 06:30:10 opie ntpd[4133]: adjusting local clock by -91.931803s
> Nov 14 06:34:22 opie ntpd[4133]: adjusting local clock by -90.983786s

ntpd does not immediately set the time, but uses adjtime(2) instead. So
your time is getting closer and closer to being accurate.

man ntpd
man adjtime

You might be interested in the -s switch of ntpd, which is set by
default by rc(8).

/Alexander

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Moritz Grimm
In reply to this post by J Moore
J Moore wrote:
> I just installed 3.8 on a Soekris net4801 that's been laying around for
> a while (unused, unpowered). I noticed after install that time was off
> by like 5 months, so I set it to within a few minutes of current
> time/date from the wall clock.
>
> I've been checking the logs, and this is what I'm seeing... this has
> been going on for about 8 hours now. Why is ntpd having to make 60+
> second adjustments every 3-5 minutes? It would appear the clock on the
> Soekris is really BFU.
[...]
> Nov 14 06:30:10 opie ntpd[4133]: adjusting local clock by -91.931803s
> Nov 14 06:34:22 opie ntpd[4133]: adjusting local clock by -90.983786s
[...]
> Nov 14 08:24:20 opie ntpd[4133]: adjusting local clock by -64.294286s
> Nov 14 08:27:59 opie ntpd[4133]: adjusting local clock by -63.612736s

OpenNTPd is working as expected. It is using adjtime(2) to skew the
clock, not set it -- in your case, it is slowing it down until it is synced.

Run rdate(8) to speed up the syncing process the hard way (the clock
will jump.) Read up on ntpd(8)'s parameter `-s' in case you ever need to
set a clock that is way off.


Moritz

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Moritz Grimm
In reply to this post by Alexander Hall
Alexander Hall wrote:
> You might be interested in the -s switch of ntpd, which is set by
> default by rc(8).

Not any longer. It was removed again to not tempt people to interrupt
the booting process via CTRL+C in case it hangs for the one or other
reason. It's easy to add back to ntpd_flags in rc.conf.local, though,
for those who want it.


Moritz

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Alexander Hall
Moritz Grimm wrote:
> Alexander Hall wrote:
>
>> You might be interested in the -s switch of ntpd, which is set by
>> default by rc(8).
>
> Not any longer. It was removed again to not tempt people to interrupt
> the booting process via CTRL+C in case it hangs for the one or other
> reason. It's easy to add back to ntpd_flags in rc.conf.local, though,
> for those who want it.

Ah, ok. Glad to hear that actually; I always thought it seemed a little
strange to have it set by default when that is not the default behaviour
of the daemon itself. I have always set ``-s'' in rc.conf.local anyway. :-)

/Alexander

Reply | Threaded
Open this post in threaded view
|

binding network interfaces

poncenby smythe
In reply to this post by J Moore
I have been googling for around ten minutes trying to find a howto  
for binding two interfaces together.
I vaguely remember doing this with linux which involved putting some  
bond statement in /etc/sysconfig/network-scripts/somefile.

The trunk command seems to do what I am looking for but I am slightly  
confused..

trunk(4):

Create a simple round robin trunk with two bge(4) Gigabit Ethernet  
interfaces:

            # ifconfig bge0 up
            # ifconfig bge1 up
            # ifconfig trunk0 trunkport bge0 trunkport bge1 \
                    192.168.1.1 netmask 255.255.255.0

i don't have a box to try this on at the moment so would trunk be  
suitable for an openbsd box which has the transmit traffic being  
spouted to em0 and receive traffic spouted to em1?
i guess it would be:

            # ifconfig em0 up
            # ifconfig em1 up
            # ifconfig trunk0 trunkport em0 trunkport em1 \
                    192.168.1.1 netmask 255.255.255.0

and what is a round-robin trunk?

thanks for any help and apologies for asking questions I could work  
out myself , I just don't have access to any openbsd at the moment.

poncenby

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

J Moore
In reply to this post by Moritz Grimm
On Mon, Nov 14, 2005 at 03:55:21PM +0100, the unit calling itself Moritz Grimm wrote:

> >I just installed 3.8 on a Soekris net4801 that's been laying around for
> >a while (unused, unpowered). I noticed after install that time was off
> >by like 5 months, so I set it to within a few minutes of current
> >time/date from the wall clock.
> >
> >I've been checking the logs, and this is what I'm seeing... this has
> >been going on for about 8 hours now. Why is ntpd having to make 60+
> >second adjustments every 3-5 minutes? It would appear the clock on the
> >Soekris is really BFU.
> [...]
> >Nov 14 06:30:10 opie ntpd[4133]: adjusting local clock by -91.931803s
> >Nov 14 06:34:22 opie ntpd[4133]: adjusting local clock by -90.983786s
> [...]
> >Nov 14 08:24:20 opie ntpd[4133]: adjusting local clock by -64.294286s
> >Nov 14 08:27:59 opie ntpd[4133]: adjusting local clock by -63.612736s
>
> OpenNTPd is working as expected. It is using adjtime(2) to skew the
> clock, not set it -- in your case, it is slowing it down until it is synced.

Hmmm... OK - I read man for adjtime(2), and I appreciate your
explanation with skewing vs setting. However, the output says
"adjusting local clock by XX s"... that seems pretty straightforward
to me. I think the output is just misleading; it should say:
"adjusting local clock to reduce error of XXs"

If you told someone you were adjusting their clock by 60 seconds, I
think most English-speaking people would conclude that you just changed
the clock by a value of 60 seconds - not that I am making an incremental
adjustment in your clock to reduce the amount of error such that it is
less than 60 seconds.

Some additional entries from my log show that in fact ntpd did finally
reach closure as of 12:57:40 today (yeah!)

Nov 14 12:29:50 opie ntpd[4133]: adjusting local clock by -5.494670s
Nov 14 12:34:05 opie ntpd[4133]: adjusting local clock by -4.846304s
Nov 14 12:37:03 opie ntpd[4133]: adjusting local clock by -3.931653s
Nov 14 12:39:18 opie ntpd[4133]: adjusting local clock by -3.201504s
Nov 14 12:43:19 opie ntpd[4133]: adjusting local clock by -2.477988s
Nov 14 12:46:12 opie ntpd[4133]: adjusting local clock by -1.698618s
Nov 14 12:49:53 opie ntpd[4133]: adjusting local clock by -0.904406s
Nov 14 12:52:53 opie ntpd[4133]: adjusting local clock by -0.227882s
Nov 14 12:57:40 opie ntpd[4133]: adjusting local clock by 0.264980s
Nov 14 12:57:40 opie ntpd[14058]: clock is now synced
Nov 14 12:59:30 opie ntpd[14058]: peer 66.11.161.129 now invalid
Nov 14 13:00:48 opie ntpd[4133]: adjusting local clock by 0.652100s
Nov 14 13:05:04 opie ntpd[4133]: adjusting local clock by -0.271403s

This is all very cool, but I still think the log messages are
misleading.
 
> Run rdate(8) to speed up the syncing process the hard way (the clock
> will jump.) Read up on ntpd(8)'s parameter `-s' in case you ever need to
> set a clock that is way off.

Thanks for that... as it turns out, the wall clock that I referred to
when I set the time using 'date' (immediately after the install) was off
"real time" by about 120 seconds... ntpd took approximately 12 hours to
work this off. This is fine - the log messages just threw me off.

Thanks,
Jay

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Alexander Hall
J Moore wrote:

>>OpenNTPd is working as expected. It is using adjtime(2) to skew the
>>clock, not set it -- in your case, it is slowing it down until it is synced.
>
>
> Hmmm... OK - I read man for adjtime(2), and I appreciate your
> explanation with skewing vs setting. However, the output says
> "adjusting local clock by XX s"... that seems pretty straightforward
> to me. I think the output is just misleading; it should say:
> "adjusting local clock to reduce error of XXs"
>
> If you told someone you were adjusting their clock by 60 seconds, I
> think most English-speaking people would conclude that you just changed
> the clock by a value of 60 seconds - not that I am making an incremental
> adjustment in your clock to reduce the amount of error such that it is
> less than 60 seconds.

It is changing it by 60 seconds, but very slowly. While still adjusting
(remember, slowly), the time is checked again, and readjusted (slowly).

http://marc.theaimsgroup.com/?l=openbsd-misc&m=110201623230329&w=2

Not really worth discussing again, since Henning voted ``no'' (kinda).

/Alexander

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

J Moore
On Mon, Nov 14, 2005 at 09:56:38PM +0100, the unit calling itself Alexander Hall wrote:

> J Moore wrote:
> >>OpenNTPd is working as expected. It is using adjtime(2) to skew the
> >>clock, not set it -- in your case, it is slowing it down until it is
> >>synced.
> >
> >
> >Hmmm... OK - I read man for adjtime(2), and I appreciate your
> >explanation with skewing vs setting. However, the output says
> >"adjusting local clock by XX s"... that seems pretty straightforward
> >to me. I think the output is just misleading; it should say:
> >"adjusting local clock to reduce error of XXs"
> >
> >If you told someone you were adjusting their clock by 60 seconds, I
> >think most English-speaking people would conclude that you just changed
> >the clock by a value of 60 seconds - not that I am making an incremental
> >adjustment in your clock to reduce the amount of error such that it is
> >less than 60 seconds.
>
> It is changing it by 60 seconds, but very slowly. While still adjusting
> (remember, slowly), the time is checked again, and readjusted (slowly).
>
> http://marc.theaimsgroup.com/?l=openbsd-misc&m=110201623230329&w=2
>
> Not really worth discussing again, since Henning voted ``no'' (kinda).

Prior discussions notwithstanding, the fact is that the log messages are
misleading. I *understand* now... if the log messages were written
differently, I never would've had to ask.

I don't know who Henning is, and I don't know what he voted "no" to, but
if he voted against a clear log message, then he voted "yes" to
confusion.

Jay

Reply | Threaded
Open this post in threaded view
|

Re: binding network interfaces

poncenby smythe
In reply to this post by poncenby smythe
On 14 Nov 2005, at 20:26, Theo de Raadt wrote:


> you probably want to use 'bridge'
>

thanks for the pointer, however I am not sure bridge is what I am  
after, let me try and explain what I meant slightly more clearly (if  
anyone is interested).....

My openbsd box has three interfaces, em2 will be used for remote access.
em0 and em1 will be used for traffic passively collected from the  
network using an ethernet tap.
so for example em0 will see all the receive traffic and em1 will see  
all the transmit traffic.

so in order to do some traffic analysis I guess it is desirable for  
these two interfaces (em0,em1) to be represented by one 'virtual'  
interface aliased to em0 and em1, so the full-duplex traffic is seen  
by tools like tcpdump/argus etc.

any thoughts? and what is a round-robin trunk?!?!?

thanks for your time

poncenby

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Matthias Kilian
In reply to this post by J Moore
On Mon, Nov 14, 2005 at 04:42:46PM -0600, J Moore wrote:
> I don't know who Henning is, and I don't know what he voted "no" to, but
> if he voted against a clear log message, then he voted "yes" to
> confusion.

Just cvs log on /usr/src/usr.sbin/ntpd and you'll learn who Henning
is (wrt ntpd).

Ciao,
        Kili

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Stan-18
On Tue, Nov 15, 2005 at 12:08:31AM +0100, Matthias Kilian wrote:
> On Mon, Nov 14, 2005 at 04:42:46PM -0600, J Moore wrote:
> > I don't know who Henning is, and I don't know what he voted "no" to, but
> > if he voted against a clear log message, then he voted "yes" to
> > confusion.
>
> Just cvs log on /usr/src/usr.sbin/ntpd and you'll learn who Henning
> is (wrt ntpd).

OK, so what did he vote no on?


--
U.S. Encouraged by Vietnam Vote - Officials Cite 83% Turnout Despite Vietcong Terror
- New York Times 9/3/1967

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Shane J Pearson
In reply to this post by J Moore
J,

On 15/11/2005, at 9:42 AM, J Moore wrote:

> Prior discussions notwithstanding, the fact is that the log  
> messages are
> misleading. I *understand* now... if the log messages were written
> differently, I never would've had to ask.

Reasonable person scenario:

o Notice odd ntpd log entries.
o #man ntpd
o Notice SECOND paragraph says:

     "ntpd uses the adjtime(2) system call to correct the local system
     time without causing time jumps.  Adjustments larger than 128ms
     are logged using syslog(3).  The threshold value is chosen to
     avoid having local clock drift thrash the log files."

o Crisis averted.

> I don't know who Henning is, and I don't know what he voted "no"  
> to, but
> if he voted against a clear log message, then he voted "yes" to
> confusion.

Come on. You've been haunting these lists for long enough to know who
Henning is. Cut the theatrics.


Shane J Pearson        shanejp netspace net au                       ->|

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

J Moore
On Tue, Nov 15, 2005 at 01:34:27PM +1100, the unit calling itself Shane J Pearson wrote:

> J,
>
> On 15/11/2005, at 9:42 AM, J Moore wrote:
>
> >Prior discussions notwithstanding, the fact is that the log  
> >messages are
> >misleading. I *understand* now... if the log messages were written
> >differently, I never would've had to ask.
>
> Reasonable person scenario:
>
> o Notice odd ntpd log entries.
> o #man ntpd
> o Notice SECOND paragraph says:
>
>     "ntpd uses the adjtime(2) system call to correct the local system
>     time without causing time jumps.  Adjustments larger than 128ms
>     are logged using syslog(3).  The threshold value is chosen to
>     avoid having local clock drift thrash the log files."
>
> o Crisis averted.

Come on, Shane - did you ever take a friggin' course in English? Are you
telling me that the passage above makes the following one-liner clear:

        'adjusting local clock by XXs'

The word 'by' is a preposition with a specific meaning in the context of
its use... it means "in the amount of"... but that's not what it means
here, is it? No, it does not. Therefore, the log entry is *inaccurate*.

So there is no crisis to be averted here, Shane. A developer for whom
English is probably a second language mis-used the language, and created
confusion. OK, that's cool... but what escapes me is this comment:

> http://marc.theaimsgroup.com/?l=openbsd-misc&m=110202770603752&w=2
> ---
> so I'll change
>       log_info("adjusting local clock by %fs", d);
> into
>       log_info("now kindly asking the kernel to adjust the clock "
>           "by %f seconds but it will not do so at once so maybe "
>           "it takes a while", d);"
> ---

I gather that he rejected the consensus that his choice of words was
confusing?

> >I don't know who Henning is, and I don't know what he voted "no"  
> >to, but
> >if he voted against a clear log message, then he voted "yes" to
> >confusion.
>
> Come on. You've been haunting these lists for long enough to know who
> Henning is. Cut the theatrics.

No theatrics intended - since the OP, I've been informed that Henning
wrote some or all of the code in ntpd. That's great - I love the code, I
think his English needs some work.

Jay

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Shane J Pearson
J,

On 15/11/2005, at 3:16 PM, J Moore wrote:

>>     "ntpd uses the adjtime(2) system call to correct the local system
>>     time without causing time jumps.  Adjustments larger than 128ms


> Come on, Shane - did you ever take a friggin' course in English?  
> Are you
> telling me that the passage above makes the following one-liner clear:
>
>     'adjusting local clock by XXs'

The "without causing time jumps" explains what you could see in the log
pretty clearly, I would have thought.


"then he voted "yes" to confusion."

This is theatrics like I'd expect from a political campaign. I almost
expect you to end it with "Vote [1] J Moore".


Shane J Pearson        shanejp netspace net au                       ->|

Reply | Threaded
Open this post in threaded view
|

Re: binding network interfaces

Alexandre Anriot
In reply to this post by poncenby smythe
> On 14 Nov 2005, at 20:26, Theo de Raadt wrote:
>
>
> > you probably want to use 'bridge'
> >
>
> thanks for the pointer, however I am not sure bridge is what I am  
> after, let me try and explain what I meant slightly more clearly (if  
> anyone is interested).....
>
> My openbsd box has three interfaces, em2 will be used for remote access.
> em0 and em1 will be used for traffic passively collected from the  
> network using an ethernet tap.
> so for example em0 will see all the receive traffic and em1 will see  
> all the transmit traffic.
>
> so in order to do some traffic analysis I guess it is desirable for  
> these two interfaces (em0,em1) to be represented by one 'virtual'  
> interface aliased to em0 and em1, so the full-duplex traffic is seen  
> by tools like tcpdump/argus etc.
>
> any thoughts? and what is a round-robin trunk?!?!?

It means that if you create a trunk interface with em0 and em1 as
trunkports, packets will be distributed between the two adapters, one
time on em0, another one on em1 (round-robin algorithm).

See trunk(4) for details.

> thanks for your time
>
> poncenby

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Henning Brauer-2
In reply to this post by J Moore
* J Moore <[hidden email]> [2005-11-15 05:19]:

> On Tue, Nov 15, 2005 at 01:34:27PM +1100, the unit calling itself Shane J Pearson wrote:
> > J,
> >
> > On 15/11/2005, at 9:42 AM, J Moore wrote:
> >
> > >Prior discussions notwithstanding, the fact is that the log  
> > >messages are
> > >misleading. I *understand* now... if the log messages were written
> > >differently, I never would've had to ask.
> >
> > Reasonable person scenario:
> >
> > o Notice odd ntpd log entries.
> > o #man ntpd
> > o Notice SECOND paragraph says:
> >
> >     "ntpd uses the adjtime(2) system call to correct the local system
> >     time without causing time jumps.  Adjustments larger than 128ms
> >     are logged using syslog(3).  The threshold value is chosen to
> >     avoid having local clock drift thrash the log files."
> >
> > o Crisis averted.
>
> Come on, Shane - did you ever take a friggin' course in English? Are you
> telling me that the passage above makes the following one-liner clear:
>
> 'adjusting local clock by XXs'
>
> The word 'by' is a preposition with a specific meaning in the context of
> its use... it means "in the amount of"... but that's not what it means
> here, is it? No, it does not. Therefore, the log entry is *inaccurate*.

it is perfectly accurate. it says "adjusting by", and that is what it
does.
it does not say "hard setting" or anything.
I won't change the log message, case closed.

> So there is no crisis to be averted here, Shane. A developer for whom
> English is probably a second language mis-used the language, and created
> confusion. OK, that's cool... but what escapes me is this comment:
>
> > http://marc.theaimsgroup.com/?l=openbsd-misc&m=110202770603752&w=2
> > ---
> > so I'll change
> >       log_info("adjusting local clock by %fs", d);
> > into
> >       log_info("now kindly asking the kernel to adjust the clock "
> >           "by %f seconds but it will not do so at once so maybe "
> >           "it takes a while", d);"
> > ---
>
> I gather that he rejected the consensus that his choice of words was
> confusing?
>
> > >I don't know who Henning is, and I don't know what he voted "no"  
> > >to, but
> > >if he voted against a clear log message, then he voted "yes" to
> > >confusion.
> >
> > Come on. You've been haunting these lists for long enough to know who
> > Henning is. Cut the theatrics.
>
> No theatrics intended - since the OP, I've been informed that Henning
> wrote some or all of the code in ntpd. That's great - I love the code, I
> think his English needs some work.

really.

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Rod.. Whitworth
On Tue, 15 Nov 2005 10:23:00 +0100, Henning Brauer wrote:

>> No theatrics intended - since the OP, I've been informed that Henning
>> wrote some or all of the code in ntpd. That's great - I love the code, I
>> think his English needs some work.
>
>really.

Apart from being a cereal (RS232) killer, no - not really!
:-))

From the land "down under": Australia.
Do we look <umop apisdn> from up over?

Do NOT CC me - I am subscribed to the list.
Replies to the sender address will fail except from the list-server.

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

Jason McIntyre-2
On Tue, Nov 15, 2005 at 08:46:07PM +1100, Rod.. Whitworth wrote:

> On Tue, 15 Nov 2005 10:23:00 +0100, Henning Brauer wrote:
>
> >> No theatrics intended - since the OP, I've been informed that Henning
> >> wrote some or all of the code in ntpd. That's great - I love the code, I
> >> think his English needs some work.
> >
> >really.
>
> Apart from being a cereal (RS232) killer, no - not really!
> :-))
>

what needs work is the poster's manners.
jmc

Reply | Threaded
Open this post in threaded view
|

Re: timekeeping on Soekris net4801 w/ ntpd. 3.8

knitti
In reply to this post by J Moore
On 11/15/05, J Moore <[hidden email]> wrote:
> Come on, Shane - did you ever take a friggin' course in English? Are you
> telling me that the passage above makes the following one-liner clear:
>
>         'adjusting local clock by XXs'
>
> The word 'by' is a preposition with a specific meaning in the context of
> its use... it means "in the amount of"... but that's not what it means
> here, is it? No, it does not. Therefore, the log entry is *inaccurate*.

come on, if you are a native speaker (as if that mattered), and fluent
_reader_, you'd have noticed the words of Alexander Hall in this thread:
"It is changing it by 60 seconds, but very slowly."
That means: the clock is changed by the amount of 60 sec, but not
immediately. Which is fine in my book, because I don't like time jumps
(they are soo confusing). And it should be correct english, which, of course,
I am not to judge, since my last friggin' course in English was back in the
days I went to school.


--knitti


>
> So there is no crisis to be averted here, Shane. A developer for whom
> English is probably a second language mis-used the language, and created
> confusion. OK, that's cool... but what escapes me is this comment:
>
> > http://marc.theaimsgroup.com/?l=openbsd-misc&m=110202770603752&w=2
> > ---
> > so I'll change
> >       log_info("adjusting local clock by %fs", d);
> > into
> >       log_info("now kindly asking the kernel to adjust the clock "
> >           "by %f seconds but it will not do so at once so maybe "
> >           "it takes a while", d);"
> > ---
>
> I gather that he rejected the consensus that his choice of words was
> confusing?
>
> > >I don't know who Henning is, and I don't know what he voted "no"
> > >to, but
> > >if he voted against a clear log message, then he voted "yes" to
> > >confusion.
> >
> > Come on. You've been haunting these lists for long enough to know who
> > Henning is. Cut the theatrics.
>
> No theatrics intended - since the OP, I've been informed that Henning
> wrote some or all of the code in ntpd. That's great - I love the code, I
> think his English needs some work.
>
> Jay

123