NTP fails synchronization with server at startup
Nigel Henry
cave.dnb at tiscali.fr
Wed Dec 5 17:17:33 UTC 2007
On Wednesday 05 December 2007 15:24, Paul Smith wrote:
> On Dec 5, 2007 2:09 PM, Mikkel L. Ellertson <mikkel at infinity-ltd.com> wrote:
> > > Yes, Craig, I did reboot the computer after having turned off ipv6,
> > > but it did not solve the problem. It is rather strange that
> > > '/sbin/service ntpd restart' works perfectly and that NTP cannot
> > > synchronize with the server at booting.
> >
> > This is not strange at all. When it tries to connect at boot, it can
> > not make the connection. When the network is up, it can make the
> > connection. I don't remember for sure, but I believe you were
> > getting a DNS error. This could be because you couldn't reach the
> > name server, or because you are running a local name server, and it
> > isn't up when NTP starts. I would also expect to see this if you
> > were using a wireless connection and network manager configured to
> > make the connection when the user logs in. (User supplied wireless
> > key instead of system supplied wireless key.)
>
> I understand your explanation. I do not use any wireless connection to
> the Internet. Moreover, my Ethernet device is configure to be
> activated when computer starts. What else should I look at?
>
> Paul
Hi Paul. One thing I would consider is to add some Internet time servers to
your /etc/ntp.conf. It is suggested that you have at least 3. Some
timeservers can be unreliable, or not even accessable at times, which may
well cause problems if you only have one. Make sure you use stratum 2, or 3,
not stratum 1 servers.
Another thing. Are you sure that ntp isn't doing it's stuff, even though the
bootup shows a fail. Before you do the ntpd restart, run the following as
user.
/usr/sbin/ntpq
then type pe, which will give you some info on which servers ntp is trying to
connect to, and how successfull it is being. You can keep typing pe at
intervals, which will show ntp's progress at reaching a point where a time
server is being used as a "sys peer". The server being used will be prefixed
by a "*". Other useable servers will be prefixed by a "+" "candidat". To quit
ntpq type q.
As a matter of course after receiving advice from the ntp list, I comment out
all the uncommented lines in /etc/ntp.conf, except the driftfile line,
including commenting out any server lines, as "I "select the servers I want
to use. Many of the uncommented lines are not needed, and in the case of the
restrict lines, can cause additional problems in getting ntp to work, unless
you know what you are doing.
I have had problems when I first started to use ntp. I have a few distro's
that run on the machine that gets it's time from the Internet. All use the
same 3 Internet servers. I still use FC2 for day to day use on this machine,
and do have a problem if my dialup connection isn't up when I boot FC2. The
ntp daemon dies. I have a script though that runs from rc.local. This pings a
webserver, and when the Internet connection is up, receives a response from
the ping. Then the script starts the ntp daemon, and all is ok.
My F8 install is on the other machine, and gets it's time from, in this case,
the machine that FC2 is running on. There are only 2 lines uncommented on
F8's /etc/ntp.conf. The driftfile line, and a server line, which happens to
be 192.168.0.230, and is pointing to the IP address of FC2 on the LAN.
Just for info I post a bit of output from ntpq> pe from FC2, that gets it's
time from the Internet.
/usr/sbin/ntpq
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
-crush.brunom.ne 88.191.16.241 3 u 35 256 377 138.418 36.829 21.347
+sd-881.dedibox. 193.52.184.106 2 u 38 256 377 137.545 4.407 10.555
*dedi.drazzib.co 193.49.205.17 2 u 40 256 377 145.066 10.910 24.633
+syrte8.obspm.fr 134.157.254.19 2 u 104 256 377 159.674 3.336 61.051
+ns1.kamino.fr 193.52.184.106 2 u 40 256 377 136.663 5.719 37.619
+ntp2.belbone.be 195.13.23.6 2 u 44 256 377 158.059 2.627 25.585
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
-crush.brunom.ne 88.191.16.241 3 u 224 256 377 138.418 36.829 21.347
+sd-881.dedibox. 193.52.184.106 2 u 227 256 377 137.545 4.407 10.555
*dedi.drazzib.co 193.49.205.17 2 u 230 256 377 145.066 10.910 24.633
+syrte8.obspm.fr 134.157.254.19 2 u 36 256 377 151.351 8.291 50.905
+ns1.kamino.fr 193.52.184.106 2 u 230 256 377 136.663 5.719 37.619
+ntp2.belbone.be 195.13.23.6 2 u 233 256 377 158.059 2.627 25.585
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
-crush.brunom.ne 88.191.16.241 3 u 232 256 377 138.418 36.829 38.303
+sd-881.dedibox. 193.52.184.106 2 u 236 256 377 137.545 4.407 28.083
*dedi.drazzib.co 193.49.205.17 2 u 238 256 377 145.066 10.910 26.332
+syrte8.obspm.fr 195.220.94.163 2 u 42 256 377 151.351 8.291 48.079
+ns1.kamino.fr 193.52.184.106 2 u 239 256 377 136.663 5.719 37.583
+ntp2.belbone.be 195.13.23.6 2 u 240 256 377 158.059 2.627 25.221
ntpq>
And a bit of output from ntpq> pe on F8
/usr/sbin/ntpq
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*192.168.0.230 88.191.12.200 3 u 106 256 377 0.511 -1.245 0.353
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*192.168.0.230 88.191.12.200 3 u 73 256 377 0.491 -0.691 0.527
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*192.168.0.230 88.191.12.200 3 u 17 256 377 0.491 -0.691 0.533
ntpq> pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*192.168.0.230 88.191.12.200 3 u 60 256 377 0.491 -0.691 0.539
ntpq>
Interestingly the refid above shows dedi.drazzib.co having an IP address
88.191.12.200, but when pinging dedi.drazzib.co from F8 at different times, I
get 2 different IP addresses, 62.210.183.19, and 62.210.183.18. This is a
good reason for using hostnames with ntp, rather than IP addresses. Hostnames
don't change, but IP addresses can.
Dont's know if any of this is usefull to you, but it's my 2¢ worth.
Nigel.
More information about the users
mailing list