Briantb pisze:
>
> Anyone have a clue?
Do you have any firewalls between this machine and the time server? Just
check if UDP/123 is not being dropped in the log. There must be
something with communication between those two.
--
Regards
Filip Kata
This is a discussion on ntpdate Command Driving Me Crazy - Aix ; Greetings! There's an appliance in the Data Center that gets its time from somewhere in space that I point two of our servers (master time servers) to for time sync. All the other servers point to these two master time ...
Greetings!
There's an appliance in the Data Center
that gets its time from somewhere in space
that I point two of our servers (master time servers)
to for time sync. All the other servers point
to these two master time servers in /etc/ntp.conf.
This works great until the appliance gets hosed.
I unplug the appliance, replug it and all is happiness again.
OK, so I'd like to be able to know when the appliance
is no longer serving time and I thought I'd do the following
on one of our master time servers:
stopsrc -s xntpd
ntpdate timeApplianceInDataCenter
rc=$?
if [[ $rc -ne 0 ]
then
# send some error message
fi
startsrc -s xntpd
However, when I run the ntpdate command I continue to get
"no server suitable for synchronization found" messages.
Here's what I know:
1. The Appliance is defined in /etc/hosts, is included in
/etc/ntp.conf, and works just fine when xntpd is running.
(I use copy/paste to ensure the proper host name is being used
for the ntpdate command.)
2. When xntpd is turned off, the "ntpdate timeApplianceInDataCenter"
command fails.
Anyone have a clue?
Briantb pisze:
>
> Anyone have a clue?
Do you have any firewalls between this machine and the time server? Just
check if UDP/123 is not being dropped in the log. There must be
something with communication between those two.
--
Regards
Filip Kata
Briantb wrote:
> Greetings!
>
> There's an appliance in the Data Center
> that gets its time from somewhere in space
> that I point two of our servers (master time servers)
> to for time sync. All the other servers point
> to these two master time servers in /etc/ntp.conf.
> This works great until the appliance gets hosed.
> I unplug the appliance, replug it and all is happiness again.
>
> OK, so I'd like to be able to know when the appliance
> is no longer serving time and I thought I'd do the following
> on one of our master time servers:
>
> stopsrc -s xntpd
> ntpdate timeApplianceInDataCenter
> rc=$?
> if [[ $rc -ne 0 ]
> then
> # send some error message
> fi
> startsrc -s xntpd
>
> However, when I run the ntpdate command I continue to get
> "no server suitable for synchronization found" messages.
>
> Here's what I know:
>
> 1. The Appliance is defined in /etc/hosts, is included in
> /etc/ntp.conf, and works just fine when xntpd is running.
> (I use copy/paste to ensure the proper host name is being used
> for the ntpdate command.)
>
> 2. When xntpd is turned off, the "ntpdate timeApplianceInDataCenter"
> command fails.
>
> Anyone have a clue?
>
Hi!
I would change your "watchdog". If you use "ntpq -p" you get your peers
like this:
server# ntpq -p
remote refid st t when poll reach delay offset
jitter
================================================== ============================
*ntp2.belbone.be 195.13.23.6 2 u 272 1024 377 46.990 -6.663
0.179
+merlin.karpo.cz 195.113.144.201 2 u 395 1024 377 51.062 9.247
1.081
+dns.univ-lyon1. 195.220.94.163 2 u 370 1024 377 55.322 7.964
3.293
The column reach is octal and shows a binary pattern. 377 means 11111111
so the last 8 polls the server was up & running. With grep/awk/perl/...
you could easy monitor your "timeApplianceInDataCenter" without restart
xntp.
with best regards
Dieter Stumpner