Re: [gpsd-dev] Clarifications needed for the time-service HOWTO

[Harlan Stenn]

"Gary E. Miller" writes:
> Yo Harlan!
> 
> On Tue, 22 Oct 2013 22:18:35 +0000
> Harlan Stenn <address@hidden> wrote:
> 
> > > I would say we prefer more verification than pool.ntp.org, but they
> > > have been at it a while longer.
> >=20
> > How about we check with Ask and find a way to work together on this.
> 
> The more the merrier.
>
> > One of those steps would be to point folks at a good central location
> > for this information.
> 
> Sadly (?), looks like gpsd is going to be that central location.

That would be a choice, and it's not my preferred choice.  But I'm not
the one writing these scripts, and I'm not the one choosing to commit
the results to the gpsd repo.

> > > > So some guidance to chimer auditing tools seems required.
> > >=20
> > > That hhas been discussed at length on the list.  ntptrace is the
> > > old way, mature but deprecated.  sntp is the new way, nice andd
> > > shiny but moving too fast to document.  ntpq somewhere in the
> > > middle.
> >=20
> > In what way is sntp a moving target?  That code has been stable and
> > pretty much unchanged for a while.
> 
> Well, for one, when I run sntp the results do not look like when other
> people run it.'  For example my sntp seems to give more debug info,
> yet does not show the remote stratum.  So we can not tell people to
> read the stratum with sntp if it often does not show the stratum.

What version of sntp are you using?  What version are they using?

Do you know what the -d flag does?

> Here is what I get, not exactly noob driendly:
> 
> # sntp -d 0.pool.ntp.org
> 22 Oct 15:25:20 sntp[30018]: Started sntp
> Starting to read KoD file /var/db/ntp-kod...
> 22 Oct 15:25:56 sntp[30018]: Error looking up 0.pool.ntp.org: System error
> Unable to resolve hostname(s)

What are you seeing that you do not want to see, and what are you not
seeing that you want to see?

> dagwood www # sntp -d 0.us.pool.ntp.org
> 22 Oct 15:26:05 sntp[30032]: Started sntp
> Starting to read KoD file /var/db/ntp-kod...
> sntp sendpkt: Sending packet to 23.92.18.113... Packet sent.
> sntp recvpkt: packet received from 23.92.18.113 is not authentic. Authentic=
> ation not enforced.
> sntp handle_pkt: Received 48 bytes from 23.92.18.113
> sntp offset_calculation:      t21: 0.033617            t34: -0.042188
>               delta: 0.075805  offset: -0.004285
> 2013-10-22 15:26:18.635819 (+0800) -0.00429 +/- 0.068069 secs
> sntp sendpkt: Sending packet to 72.8.140.222... Packet sent.
> sntp recvpkt: packet received from 72.8.140.222 is not authentic. Authentic=
> ation not enforced.
> sntp handle_pkt: Received 48 bytes from 72.8.140.222
> sntp offset_calculation:      t21: 0.022373            t34: -0.012774
>               delta: 0.035147  offset: 0.004800
> 2013-10-22 15:26:18.671056 (+0800) +0.004800 +/- 0.028198 secs
> sntp sendpkt: Sending packet to 108.61.73.243... Packet sent.
> sntp recvpkt: packet received from 108.61.73.243 is not authentic. Authenti=
> cation not enforced.
> sntp handle_pkt: Received 48 bytes from 108.61.73.243
> sntp offset_calculation:      t21: 0.035720            t34: -0.050951
>               delta: 0.086671  offset: -0.007615
> 2013-10-22 15:26:18.757815 (+0800) -0.007615 +/- 0.025040 secs
> sntp sendpkt: Sending packet to 198.110.48.12... Packet sent.
> sntp recvpkt: packet received from 198.110.48.12 is not authentic. Authenti=
> cation not enforced.
> sntp handle_pkt: Received 48 bytes from 198.110.48.12
> sntp offset_calculation:      t21: 0.016210            t34: -0.042004
>               delta: 0.058214  offset: -0.012897
> 2013-10-22 15:26:18.816145 (+0800) -0.012897 +/- 0.047775 secs

See my previous questions above.  Here's what I see:

address@hidden sntp 0.pool.ntp.org
sntp address@hidden Sun Nov 25 09:51:12 UTC 2012 (6)
kod_init_kod_db(): Cannot open KoD db file /var/db/ntp-kod
2013-10-22 17:13:52.389985 (+0800) +0.5167 +/- 0.065132 0.pool.ntp.org 
208.79.93.34 s3
address@hidden

... and I'm going to add a %m to the "Cannot open KoD db file..." line.

You've got the date/time and timezone offset, the correction, the
expected error, the hostname, host IP, and stratum.

> To create a document to show the user how to create a 'good' stratum 1
> server using a GPS and gpsd.  Along the way the user needs to know how
> his chimer compares to that of others and knowing that how to adjust his
> fudge.

If we are going to remove all refclocks from ntpd that gpsd can cover,
that's a strong point.  If there are GPS refclocks that ntpd currently
covers that gpsd does not, then the argument of putting these utils and
docs in gpsd is much weaker.

If a significant poriton of the docs and/or utils are generallhy useful,
that's an argument for putting this in the NTP distribution.

> > > My gut feel is that we are gonna end up writing our own tools.  :-)
> >=20
> > If they are tools for NTP would it be better if they were in the NTP
> > distribution?
> 
> Ideally, maybe yes.  But esr waits for noone.

What's to wait for?  I can turn folks in to NTP committers pretty
easily, and if they don't want to do that we also accept patches,
preferably as attachments to bug reports at bugs.ntp.org.

H