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

Till Wimmer news-dfn at substring.ch
Sat Feb 24 17:03:56 UTC 2007


Hi there,

i try to settle some time accuracy in our office environment...

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:
--- snip ---
  driftfile /var/lib/ntp/drift/ntp.drift

  server 127.127.1.0
  fudge 127.127.1.0 stratum 13

  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

  disable auth
--- snap ---

Short after starting, everythings looks ok:
      remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*typhoon.aput.ne 192.12.19.20     2 u  227  256  377  170.905  -14.600   5.776
+bitvalve.org    192.53.103.108   2 u  166  256  377   32.163  -15.190   4.488
+8.15.10.42      132.163.4.103    2 u  170  256  377  105.118  -11.706   3.857
-ns.oredin.net   193.190.230.66   2 u  179  256  277  141.835   36.960  17.438
  LOCAL(0)        LOCAL(0)        13 l   49   64  377    0.000    0.000   0.001

But after 3..5 hours all peers are rejected (ntpq> as) and LOCAL(0) is the sync source.

I checked the firewall, disabled all restrict settings, disabled chroot and running as unprivileged user, a.s.o.

Now i found that pps stability is always 512ppm and pll has a huge offset.

ntpdc> kern:
  pll offset:           4294.95 s
  pll frequency:        22.694 ppm
  maximum error:        0.364061 s
  estimated error:      0.017754 s
  status:               0001  pll
  pll time constant:    4
  precision:            1e-06 s
  frequency tolerance:  512 ppm
  pps frequency:        0.000 ppm
  pps stability:        512.000 ppm
  pps jitter:           0.0002 s
  calibration interval: 4 s
  calibration cycles:   0
  jitter exceeded:      0
  stability exceeded:   0
  calibration errors:   0

I set the hwclock -w and started fro scratch, but short after the offset was the same.
I can't believe that the clock of this server is this terribly inaccurate... IMHO it's "professional" hardware. At least it was 
expensive ;)

Any help would be very much appreciated!
Till




More information about the questions mailing list