[ntp:questions] debugging strange ntp in virtual environment
Harlan Stenn
stenn at ntp.org
Wed Sep 18 20:45:31 UTC 2013
"Riccardo Castellani" writes:
> My NTP server has strange problem and my clients cannot update their time
> clock.
> My NTP server is Linux and NTP version is 4.2.0.a
That's 10 years old, and 3 (coming up on 4) major releases ago.
http://support.ntp.org/Dev/ReleaseTimeline
> This NTP server is a virtual machine (by P2V) but few months ago it was
> working fine.
> My clients get following msg when I run ntpd daemon: 'no server suitable for
> synchronization found'
> In client 'A' I run this command 'ntpq -pn' :
>
> remote refid st t when poll reach delay offset
> disp
> =============================================================================
> =
> myntpserver 0.0.0.0 16 u 29 64 0 0.00 0.000
> 16000.0
client A is not getting any packets exchanged with myntpserver.
> In client 'B' I run this command 'ntpq -pn' :
> remote refid st t when poll reach delay offset
> disp
> =============================================================================
> =
> myntpserver 83.84.69.80 16 u 38 64 0 2.78 -758.11 16000.0
Ditto - your "reach" value is 0.
> In NTP server 'ntpq -pn' shows:
>
> remote refid st t when poll reach delay offset
> jitter
> =============================================================================
> =
> *193.204.114.232 .CTD. 1 u 22 64 375 20.863 -672.31
> 232.744
> +193.204.114.233 .CTD. 1 u 22 64 377 20.846 -884.47
> 178.639
> +193.79.237.14 .PPS. 1 u 17 64 377 26.264 -819.79
> 181.462
> x192.93.2.20 .GPSi. 1 u 16 64 377 32.996 -1039.3
> 262.130
> +131.188.3.221 .DCFp. 1 u 22 64 377 42.731 -818.20
> 179.726
Those offsets fluctuate over 360 msec - that's a lot. The jitter is
high too.
> I reset drift file, I restarted daemon but 'offset' value increases in few
> hours until to be 800-1000.
Yes, it's still trying to figure out the drift of your clock.
> In NTP server 'ntpq -c rv' shows:
>
> assID=0 status=c6f4 sync_alarm, sync_ntp, 15 events, event_peer/strat_chg,
> version="ntpd 4.2.0a at 1.1196-r Fri May 12 09:51:35 EDT 2006 (1)"?,
> processor="i686", system="Linux/2.6.11-1.1369_FC4smp", leap=11,
> stratum=16, precision=-18, rootdelay=0.000, rootdispersion=197.910,
> peer=18845, refid=STEP,
> reftime=00000000.00000000 Thu, Feb 7 2036 7:28:16.000, poll=6,
> clock=0xd5e45b57.a691d9b1, state=2, offset=0.000, frequency=0.000,
> noise=775.898, jitter=238.455, stability=0.000
>
>
> I installed another server (this time physical server) in the same lan,
> configured in the same way, same ntp version and it works fine!!!
> It's physical server, unique difference ! Clients can update time from
> physical server and no update from the virtual server !
>
> In physical server 'ntpq -pn' shows:
>
> remote refid st t when poll reach delay offset
> jitter
> =============================================================================
> =
> *193.204.114.232 .CTD. 1 u 216 512 335 22.177 -4.268
> 0.244
> +193.204.114.233 .CTD. 1 u 98 512 331 20.756 -4.873
> 0.351
> -131.188.3.221 .DCFp. 1 u 201 512 335 49.258 -12.984
> 2.276
> +193.79.237.14 .PPS. 1 u 203 512 335 27.570 -5.418
> 1.051
Much better offset and drift values. But you seem to be losing about 2
out of every 8 NTP polls, based on the "reach" numbers.
> In physical server 'ntpq -c rv' shows:
>
> assID=0 status=0644 leap_none, sync_ntp, 4 events, event_peer/strat_chg,
> version="ntpd 4.2.0a at 1.1196-r Fri May 12 09:51:35 EDT 2006 (1)"?,
> processor="i686", system="Linux/2.6.11-1.1369_FC4smp", leap=00,
> stratum=2, precision=-17, rootdelay=22.177, rootdispersion=35.682,
> peer=64524, refid=193.204.114.232,
> reftime=d5e45b36.ce35610a Wed, Sep 18 2013 19:02:46.805, poll=9,
> clock=0xd5e45c4d.f6807bbb, state=4, offset=-4.293, frequency=8.223,
> noise=0.823, jitter=0.951, stability=40.108
H
More information about the questions
mailing list