Solaris 8 Timeserver xntp error xx seconds exceeds sanity limit

From: andrea.soliva freehome.ch (andrea.soliva@freehome.ch)
Date: Fri Jan 03 2003 - 17:56:00 EST


Hi

Now I have a problem since the sommer time was changed to the winter time in
october....? I have Serial Mouse Clock for Time Sinchronisation and the
software for the synchronisation I use is xntp newest version. I compiled
the software with now problem and in the mean time all is OK and all servers
will be synchronized with the correct time. On the client I use a cron job
with following command:

/opt/xntp-4.1.1b/bin/ntpdate -v -s guiadmin

(-v for verbose and -s for loggin local in messages)

The entry which will be writen in the messages log is:

Dec 31 08:00:01 proteus ntpdate[394]: [ID 702911 daemon.notice] ntpdate
4.1.1b@1
.829 Wed Nov 27 23:43:28 MET 2002 (1)
Dec 31 08:00:01 proteus ntpdate[394]: [ID 702911 daemon.notice] adjust time
serv
er 192.168.100.9 offset 0.069116 sec

You see all OK...also the local time on the clients are OK I mean if you use
the following command the time is absolutly correct:

date (actual local time)

date -u (UTC time it means for my localisation - 1 hour = middle european
time)

On the timeserver itself I use the command:

/opt/xntp-4.1.1b/bin/ntpd -g -c /opt/xntp-4.1.1b/conf/xntp.conf -p
/var/run/xntp.pid

(-g specially for Solaris if a client will kill the main process and -c for
special conf file; -p for pid file)

In the conf file itself I used following information:

server 127.127.8.0 prefer mode 16
logfile /var/log/xntp/xntp.log
driftfile /var/log/xntp/xntp.drift

# The following peer host are client machines
# for synchronize the time:
peer < ip adress of client >

# This configuration will log all events:
#logconfig =all

# This would just list the synchronizations
# state of ntpd and the major system events.
# For a simple reference server, the following
# minimum message configuration could be useful:
#logconfig =syncstatus +sysevents

# This configuration will list all clock information
# and synchronization information. All other events
# and messages about peers, system events and so on
# is suppressed:
logconfig =syncall +clockall

( preffer mode 16 meens special code for my clock...!)

Now what is my problem.....from time to time..not linear (2 to 5 days) the
main process on the timeserver killes itself with following error message in
the log file:

 3 Jan 18:05:00 ntpd[506]: PARSE receiver #0: SYNCHRONIZED
 3 Jan 18:05:00 ntpd[506]: clock GENERIC(0) event 'clk_okay' (0x00)
 3 Jan 18:05:00 ntpd[506]: synchronisation lost
 3 Jan 18:06:11 ntpd[506]: time correction of 315619200 seconds exceeds
sanity l
imit (1000); set clock manually to the correct UTC time.

If I controll the time on the timeserver itself with date or date -u the
time is absolutly correct and it is not a "sanity limit of 1000" ??????? It
looks like the main process could not write the time in the hardware
clock....? Could it be....? What is wrong...anybody also knows such a
problem.....befor the time changed from sommer to winter time the processes
where running a year with no problem...! I tried also to debugg but there is
allways the same messages above...! I looked also in the different sources
in the internet and found something with a kernel parameter...something with
"tickad..." I don't know realliy the parameter but I know it is to use in
"/etc/system"......could this solve my problem....what I have to do....?
Write me if anybody knows something about such a phaenomen....thanks in
advanced....

kind regards

MR. Andrea Soliva
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:25:32 EDT