[Voyage-linux] Strange problem with NTPD

[Mark Powell]

Well I've spent a few hours on this over the last couple of days, and
I have completely rebuilt the server.

The NS records for my domain should be updated, so the domain should
be accessible by now.

I have updated the default ntp.conf file, so you can now do a ntpq -pn
ntp.typical.me.uk and the web server at ntp.typical.me.uk is also up
and running, plus the cron job that updates the traffic image is
working fine. I can see it in /var/log/syslog - no errors

I don't really know what the original problem was, but re-flashing
voyage onto the CF and apt-getting all the packages, ntp, nginx etc
has worked this time.

Regards

Mark

On 1 July 2012 11:12, Adrian Reyer <are at lihas.de> wrote:
> Hi Mark,
>
> On Thu, Jun 28, 2012 at 11:39:56PM +0100, Mark Powell wrote:
>> If you go to ntp.typical.me.uk you can see for yourself that it's all
>> running fine except for one major problem - ntpd doesn't run properly
>> anymore. I received a mail from the pool administrator within a few
>> hours informing that me that the server is down. If I do a ntpq -p
>> ntp.typical.me.uk from another host on my LAN, I get either a 'read
>> connection refused' or a 'request timeout' error, whereas previously I
>> could always see the peers.
>
> $ntpq -p ntp.typical.me.uk
> ntp.typical.me.uk: timed out, nothing received
>
> I guess the issue still persists.
>
>> If I do /etc/init.d/ntp stop the first time I get
>> Stopping NTP server: ntpdstart-stop-daemon: warning: failed to kill
>> 1759: No such process
>> .
>> but then I can restart, and the ntpd daemon is running, but unresponsive.
>
> ntpd takes some time till it answers as it wants to get a decent time
> first. In my eyes, the big advandtage of 2d13 over 2d3 is the added cmos
> battery, keeping the time current over restarts. Is your local time
> fine? Perhaps the battery is depleted?
> Is DNS working?
> I'd use "tcpdump -n -i ethN port 123" and check if you get the requests on
> the box at all. And more, if you start ntpd, if it does requests to the
> outside. If you have a box outside, you can try and do a tcpdump there
> as well how a port 123 udp request looks like, perhaps something like
> NAt issues showed up.
> ntpdc has some commands that might enlighten you about the status as
> well.
>
> Regards,
>         Adrian
> --
> LiHAS - Adrian Reyer - Hessenwiesenstraße 10 - D-70565 Stuttgart
> Fon: +49 (7 11) 78 28 50 90 - Fax:  +49 (7 11) 78 28 50 91
> Mail: lihas at lihas.de - Web: http://lihas.de
> Linux, Netzwerke, Consulting & Support - USt-ID: DE 227 816 626 Stuttgart