[ntp:questions] Cannot synchronize to server with local clock

Christoph Bartoschek bartoschek at gmx.de
Wed Jul 13 07:29:58 UTC 2005


Hi,

our network consists of two subsystems. One is a DCE/DFS environment and the
other one is a Linux/NFS environment. There have been several problems
because the time is different in both systems. The DCE environment
synchronizes itself via the DCE Distributed Time Service.  The linux
machines used setclk to get the time from a DCE server, but this is to
inaccurate. Now we plan to use NTP to synchronize both systems. 

According to the DCE documentation one has to start a NTP Server on a DCE
machine which does not change the local clock but uses it as a reference.
The other machines use this server to synchronize themselves.

Now I want to simulate the situation on a small testbed consisting of three
linux machines before I change our servers. However this test fails.
Reading the docs and the FAQ I came up with the following configuration for
the server:

server 127.127.1.0 prefer               
fudge  127.127.1.0 stratum 4 refid DTSS 
disable ntp

The requirements are that the local clock is used as reference and that it
is not modified. When I start the server I get the following log messages:

13 Jul 09:07:41 ntpd[14000]: system event 'event_restart' (0x01) status
'sync_alarm, sync_unspec, 1 event, event_unspec' (0xc010)
13 Jul 09:07:42 ntpd[14000]: peer LOCAL(0) event 'event_reach' (0x84) status
'unreach, conf, 1 event, event_reach' (0x8014)
13 Jul 09:10:58 ntpd[14000]: system event 'event_peer/strat_chg' (0x04)
status 'sync_alarm, sync_local_proto, 2 events, event_restart' (0xc521)
13 Jul 09:10:58 ntpd[14000]: synchronized to LOCAL(0), stratum 4


This seems ok for me. The clients however are not able to synchronize to the
server:

wilma:~ # ntpdate -d burns
13 Jul 09:25:54 ntpdate[3962]: ntpdate 4.2.0a at 1.1213-r Wed Jan 26 17:22:51
UTC 2005 (1)
Looking for host burns and service ntp
host found : burns.wesseling.pontohonk.de
transmit(192.168.0.101)
receive(192.168.0.101)
transmit(192.168.0.101)
receive(192.168.0.101)
transmit(192.168.0.101)
receive(192.168.0.101)
transmit(192.168.0.101)
receive(192.168.0.101)
transmit(192.168.0.101)
192.168.0.101: Server dropped: strata too high
server 192.168.0.101, port 123
stratum 16, precision -19, leap 11, trust 000
refid [192.168.0.101], delay 0.02574, dispersion 0.00005
transmitted 4, in filter 4
reference time:    00000000.00000000  Thu, Feb  7 2036  7:28:16.000
originate timestamp: c67f4082.c2d81ade  Wed, Jul 13 2005  9:25:54.761
transmit timestamp:  c67f4082.79b23dd5  Wed, Jul 13 2005  9:25:54.475
filter delay:  0.02673  0.02577  0.02574  0.02577
         0.00000  0.00000  0.00000  0.00000
filter offset: 0.286084 0.285616 0.285609 0.285587
         0.000000 0.000000 0.000000 0.000000
delay 0.02574, dispersion 0.00005
offset 0.285609

13 Jul 09:25:54 ntpdate[3962]: no server suitable for synchronization found


The clock seems to be dropped because the stratum is 16 and not 4.  How does
this come?

Using the rv command in  ntpq on the server I get:

assID=0 status=c524 sync_alarm, sync_local_proto, 2 events,
event_peer/strat_chg,
version="ntpd 4.2.0a at 1.1190-r Sat Mar 19 19:20:10 UTC 2005 (1)"?,
processor="i686", system="Linux/2.6.11.4-21.7-default", leap=11,
stratum=16, precision=-19, rootdelay=0.000, rootdispersion=16.920,
peer=42196, refid=INIT,
reftime=00000000.00000000  Thu, Feb  7 2036  7:28:16.000, poll=6,
clock=0xc67f40a7.a0fa69be, state=1, offset=0.000, frequency=0.000,
noise=0.002, jitter=0.002, stability=0.000

Here the stratum is also 16 and not 4.

How can I achieve that the server uses the local clock as reference and that
it is able to give the time to the clients?

Christoph




More information about the questions mailing list