Bug#393509: [pkg-ntp-maintainers] Bug#393509: ntp: does not adjust time

[Christoph Martin]

Hi Kurt,

Kurt Roeckx schrieb:
> On Wed, Oct 18, 2006 at 10:17:39AM +0200, Christoph Martin wrote:
>> I now have it running for over 12 hours with only one server to sync
>> with and I still have the same problem. I initially set the time with
>> ntpdate because I could not get ntpd to sync it:
>>
>>  134.93.226.19   134.93.227.254   3 u    6   64  377    0.416  -469257
>> 1570.32
> 
> So, this one it totaly off, and I wonder why it doesn't synchronize.

I can't get it synchronised on this special machine. It works on the
other machines with the new ntpd.

>> The next server (134.93.226.19) is running sarge version of ntpd and
>> successfully syncs:
>>
>>  134.93.226.2    131.188.3.220    2 u  199 1024    3    0.983    2.203
>>  0.310
>> *134.93.227.254  192.53.103.108   2 u  443 1024  377    0.446   -0.277
>>  0.340
> 
> This actually the first output that makes any sense so far.  Altough, I
> have to wonder why the first peer only replied to the last 2 packets.

This is because I just restartet the first peer to get a better
configuration.

>> To summerize it: 134.93.226.2 and 134.93.227.254 are firewall hosts
>> which act as ntp proxies (and others). They have the same configuration
>> and peer eachother. 134.93.226.19 is the internal network ntp server
>> which connects to the two firewall proxies. Most of the other servers
>> just connect to the internal server or additionally to the proxies.
> 
> So you have:
> 134.93.226.19 (int1) peers with:
> - 134.93.226.2 (fw1)
> - 134.93.227.254 (fw2)
> 
> fw1 and fw2 synchronize with external sources and peer with each other.
> 
> All 3 seem to be synchronize properly, atleast currently.

For some time now.

> Some other hosts has all 3 of them as server, but does not synchronize.
> 
> In your previous mails you had 134.93.226.2 (fw1) always as stratum 5,
> which doesn't make sense.
> 
> Anyway, it's recommended that you have atleast 3 peers with a higher
> stratum and then peer with each other.  fw1 and fw2 only have 2 higher
> stratum servers, peer with each other, and has a stratum 3 server too
> which is probably totaly useless, even though it might give more
> accurate time.

I tried to sanitize the complete setup. But it does not change the
behavior of the ntpds. On the one machine it is always not synching. On
the other machines the poll value is always at 64 and never rising. Also
the offset is most times at least a second and the jitter is
considerably high. The old ntp deamons react perfectly well and have low
jitters and offset (about some milliseconds).

I am now setting up another test on a machine in completely different
network outside of my firewall directly connected to the debian.pool
servers.

Please let me know if you have any more hints what to try or how to debug.

Christoph

-- 
============================================================================
Christoph Martin, Leiter der EDV der Verwaltung, Uni-Mainz, Germany
 Internet-Mail:  Christoph.Martin at Verwaltung.Uni-Mainz.DE
  Telefon: +49-6131-3926337
      Fax: +49-6131-3922856

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-ntp-maintainers/attachments/20061024/cb05a2d0/signature.pgp