ntpd does not re-query servers, when a new interface appears

[Dominic Fandrey]

On 09/03/2010 14:33, Jeremy Chadwick wrote:
> On Tue, Mar 09, 2010 at 05:30:45AM -0800, Jeremy Chadwick wrote:
>> On Tue, Mar 09, 2010 at 02:17:48PM +0100, Dominic Fandrey wrote:
>>> On 09/03/2010 11:27, Ian Smith wrote:
>>>> On Tue, 9 Mar 2010, Dominic Fandrey wrote:
>>>>  > ntpd tracks interface updates, however it does not requery
>>>>  > servers, when they occur. This was less than an hour ago,
>>>>  > at my university, the notebook boots and is not connected
>>>>  > to anything:
>>>>  > 
>>>>  >  9 Mar 08:07:17 ntpd[1510]: logging to file /var/log/ntpd
>>>>  >  9 Mar 08:07:17 ntpd[1510]: precision = 2.234 usec
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #0 wildcard, 0.0.0.0#123 Disabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #1 wildcard, ::#123 Disabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #2 bge0, 192.168.1.12#123 Enabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #3 lo0, fe80::1#123 Enabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #4 lo0, ::1#123 Enabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on interface #5 lo0, 127.0.0.1#123 Enabled
>>>>  >  9 Mar 08:07:17 ntpd[1510]: Listening on routing socket on fd #26 for interface updates
>>>>  >  9 Mar 08:07:17 ntpd[1510]: kernel time sync status 2040
>>>>  >  9 Mar 08:07:17 ntpd[1510]: frequency initialized 3.155 PPM from /var/db/ntpd.drift
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: 0.de.pool.ntp.org
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `0.de.pool.ntp.org', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: 1.de.pool.ntp.org
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `1.de.pool.ntp.org', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: 2.de.pool.ntp.org
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `2.de.pool.ntp.org', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: ntp1.rz.uni-karlsruhe.de
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `ntp1.rz.uni-karlsruhe.de', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: ntp1.rz.uni-karlsruhe.de
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `ntp1.rz.uni-karlsruhe.de', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: ntp3.rz.uni-karlsruhe.de
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `ntp3.rz.uni-karlsruhe.de', giving up on it
>>>>  >  9 Mar 08:07:20 ntpd[1542]: host name not found: ntp4.rz.uni-karlsruhe.de
>>>>  >  9 Mar 08:07:20 ntpd[1542]: couldn't resolve `ntp4.rz.uni-karlsruhe.de', giving up on it
>>>>  > 
>>>>  > So ntpd has given up on all the servers listed in the ntp.conf file.
>>>>
>>>> Yes, but it looks more like name service that's not operating, ntpd 
>>>> seems to be doing its best but can't resolve the hostnames?
>>>
>>> Why would I have named running on a notebook? This is a notebook,
>>> which is not connected to the internet.
>>>
>>>>  > I then proceed to connect to the wireless network and proceed to log
>>>>  > into two VPNs:
>>>>  > 
>>>>  >  9 Mar 08:08:58 ntpd[1510]: Listening on interface #6 wlan0, 192.168.75.58#123 Enabled
>>>>  >  9 Mar 08:09:00 ntpd[1510]: Listening on interface #7 tun0, 193.196.120.15#123 Enabled
>>>>  >  9 Mar 08:09:04 ntpd[1510]: Listening on interface #8 tun1, 141.3.162.67#123 Enabled
>>>>  > 
>>>>  > Over interface #8 some of the servers are actually available, but
>>>>  > ntpq -p still states:
>>>>  > No association ID's returned
>>>>  > 
>>>>  > Only when I restart ntpd, it operates as expected:
>>>>  >      remote           refid      st t when poll reach   delay   offset  jitter
>>>>  > ==============================================================================
>>>>  >  zit-net2.uni-pa .STEP.          16 u    -  512    0    0.000    0.000   0.000
>>>>  >  alpha.rueckgr.a .STEP.          16 u    -  512    0    0.000    0.000   0.000
>>>>  >  ntp.goneco.de   .STEP.          16 u    -  512    0    0.000    0.000   0.000
>>>>  > +proxy4.rz.uni-k 129.13.64.17     2 u   30  128  271    2.937    2.530   1.891
>>>>  > +proxy2.rz.uni-k 129.13.64.17     2 u   58  128  375    3.593   -8.981   1.837
>>>>  > *proxy1.rz.uni-k 129.13.64.17     2 u   15  128  271    3.297    8.244   1.487
>>>>
>>>> I've always had to restart named after losing / regaining an interface, 
>>>> most noticeably after a suspend/resume (eg a low battery suspend), so I 
>>>> run /etc/rc.d/named restart from rc.resume.  This looks like a similar 
>>>> issue perhaps, though I don't see why restarting only ntpd would fix it.
>>>
>>> As I said, named doesn't run at all. When the notebook gets an
>>> internet connection, ntpd recognizes this. It somehow doesn't
>>> occur to it, though, that it might be able to resolve the
>>> servers, now.
>>
>> I believe this is the problem.  Note that you'll need to add an SSL
>> cert. exception for this site due to them using self-signed certs.
>>
>> https://support.ntp.org/bugs/show_bug.cgi?id=987
> 
> Oh, and/or this one too, which mentions how to get the FORCE_DNSRETRY
> feature -- you get to rebuild ntpd.  :-)
> 
> https://support.ntp.org/bugs/show_bug.cgi?id=1178
> 
> Doesn't sound like this issue is FreeBSD-specific, so you might want to
> chime in on those NTP bugs.

OK, thanks for all the replies. I'll visit the NTP folks and
make a little noise there.

In the meantime, your comments made me realize, that I can circumvent
this problem by adding the ntp pools to my /etc/hosts file.

Regards

-- 
A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?