[ntp:questions] Stability problem on PowerEdge (SuSE 9.3)

[Till Wimmer]

hi steve,

Steve Kostecke wrote:
> On 2007-02-24, Till Wimmer <news-dfn at substring.ch> wrote:
> 
>> Our main server is running SuSE 9.3 on AMD64. I installed the xntp
>> package from SuSE (Vers. 4.2.0a-35). Hardware: DELL Power Edge 2850
>>
>> uanme -a:
>>   Linux office 2.6.11.4-21.15-smp #1 SMP Tue Nov 28 13:39:58 UTC 2006 x86_64 x86_64 x86_64 GNU/Linux
> 
>> /etc/ntp.conf:
> 
>>   server 127.127.1.0
>>   fudge 127.127.1.0 stratum 13
> 
> You should comment out the LocalCLK until you fix your stability
> problem. And then, you should only re-enable the LocalCLK if you
> understand why you need it.
Well i confess, i'm new to ntp problematics...  i thought we need this because all other servers in the LAN are sync'ed to this 
one. We need, at least, the same time on every server.

> 
>>   server 0.debian.pool.ntp.org iburst
>>   server 1.debian.pool.ntp.org iburst
>>   server 2.debian.pool.ntp.org iburst
>>   server 3.debian.pool.ntp.org iburst
> 
> Some people will tell you that this is a poor choice of servers (e.g.
> due to excessive delay). That may be true but it's not germane to your
> problem.
I changed them to ch.pools.... see follow ups
> 
>>   disable auth
> 
> Disabling auth is not usually a good idea because it can allow someone
> with ntpdc to tinker with your ntpd settings remotely. BUT ... this has
> nothing to do with your problem.
i disabled everything that could be restrective in some way.
> 
>> Short after starting, everythings looks ok:
> 
> <snip>
> 
>> But after 3..5 hours all peers are rejected (ntpq> as) and LOCAL(0) is
>> the sync source.
> 
> We need to see 'ntpq -p' after your ntpd has been running for those 3-5
> hours. We also need to see any ntpd messages in your syslog for that
> same period.

Tue Feb 27 16:29:09 CET 2007:
      remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
  entry.verboten. 130.149.17.21    2 u  296 1024  377  2110.41  906.197  30.261
  ns1.nexellent.n 193.190.230.66   2 u   36 1024  377  2021.64  867.787   6.034
  bryan.solnet.ch 192.53.103.104   2 u 1005 1024  377  2015.45  878.894   2.276
  idaixds2.unizh. 130.60.128.7     3 u 1027 1024  376  2012.79  877.746   2.344
*LOCAL(0)        LOCAL(0)        10 l    7   64  377    0.000    0.000   0.001

/var/log/ntp:
26 Feb 22:01:28 ntpd[26518]: synchronized to 217.147.208.1, stratum 2
26 Feb 22:25:17 ntpd[26518]: synchronized to 80.74.132.178, stratum 2
26 Feb 22:52:34 ntpd[26518]: offset -0.004164 sec freq 10.024 ppm error 0.007587 poll 8
26 Feb 23:18:30 ntpd[26518]: synchronized to 217.147.208.1, stratum 2
26 Feb 23:52:37 ntpd[26518]: offset -0.004624 sec freq 9.597 ppm error 0.000965 poll 9
27 Feb 00:06:12 ntpd[26518]: synchronized to 80.74.132.178, stratum 2
27 Feb 00:10:00 ntpd[26518]: synchronized to 212.101.4.253, stratum 3
27 Feb 00:15:52 ntpd[26518]: synchronized to LOCAL(0), stratum 10
27 Feb 00:52:40 ntpd[26518]: offset 0.000000 sec freq 9.597 ppm error 0.000001 poll 10
27 Feb 01:52:43 ntpd[26518]: offset 0.000000 sec freq 9.597 ppm error 0.000001 poll 10
27 Feb 02:52:46 ntpd[26518]: offset 0.000000 sec freq 9.597 ppm error 0.000001 poll 10
-- snap --
last line stays the same for the remaining of the day


> 
>> Now i found that pps stability is always 512ppm and pll has a huge offset.
> 
> You're not using PPS. Look at the PLL frequency, not the PPS stability.
i got this... see follow ups
> 
> 
so the question remains...

bg
till