Can't getCorrect Time in Fedora 8

[Jim]

Fedora 8 won't keep correct time.
"Unable to connect to time server"  , it doesn' make any difference what 
time server you try to connect to.
"ntpd" is running in services.

[Mikkel]

Jim wrote:
> Fedora 8 won't keep correct time.
> "Unable to connect to time server"  , it doesn' make any difference what
> time server you try to connect to.
> "ntpd" is running in services.

Dumb question time:

To you have a full-time internet connection?

Is your network configured properly?
(Can you connect to web sites like http://www/google.com?)

What does /var/log/messages show? (nptd messages)

Mikkel

[Jim]

Mikkel L. Ellertson wrote:
> Jim wrote:
> > Fedora 8 won't keep correct time.
> > "Unable to connect to time server"  , it doesn' make any difference what
> > time server you try to connect to.
> > "ntpd" is running in services.
>
> Dumb question time:
>
> To you have a full-time internet connection?

yes
> Is your network configured properly?
> (Can you connect to web sites like http://www/google.com?)

yes
> What does /var/log/messages show? (nptd messages)
>
> Mikkel

ntpd[1906]: Listening on interface #2 eth0, fe80::20f:eaff:fe8c:135c#123 
Enabled
Dec 23 14:15:42 x86_64 ntpd[1906]: Listening on interface #3 eth0, 
172.16.1.38#123 Enabled

Dec 21 12:34:59 x86_64 ntpd[1891]: ntpd exiting on signal 15

Dec 21 12:41:34 x86_64 ntpd[1905]: ntpd 4.2.4p4@1.1520-o Tue Apr 15 
14:22:43 UTC 2008 (1)
Dec 21 12:41:34 x86_64 ntpd[1906]: precision = 1.000 usec
Dec 21 12:41:34 x86_64 ntpd[1906]: Listening on interface #0 wildcard, 
0.0.0.0#123 Disabled
Dec 21 12:41:34 x86_64 ntpd[1906]: Listening on interface #1 wildcard, 
::#123 Disabled
Dec 21 12:41:34 x86_64 ntpd[1906]: kernel time sync status 0040
Dec 21 12:41:34 x86_64 ntpd[1906]: frequency initialized 0.000 PPM from 
/var/lib/ntp/drift
Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "0.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "1.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "2.fedora.pool.ntp.org" 
invalid host address, ignored

Dec 21 12:56:34 x86_64 ntpd[1906]: ntpd exiting on signal 15

Dec 21 12:59:33 x86_64 ntpd[1904]: ntpd 4.2.4p4@1.1520-o Tue Apr 15 
14:22:43 UTC 2008 (1)
Dec 21 12:59:33 x86_64 ntpd[1905]: precision = 1.000 usec
Dec 21 12:59:33 x86_64 ntpd[1905]: Listening on interface #0 wildcard, 
0.0.0.0#123 Disabled
Dec 21 12:59:33 x86_64 ntpd[1905]: Listening on interface #1 wildcard, 
::#123 Disabled
Dec 21 12:59:33 x86_64 ntpd[1905]: kernel time sync status 0040
Dec 21 12:59:33 x86_64 ntpd[1905]: frequency initialized 0.000 PPM from 
/var/lib/ntp/drift
Dec 21 12:59:33 x86_64 ntpd[1905]: getaddrinfo: "0.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 21 12:59:33 x86_64 ntpd[1905]: getaddrinfo: "1.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 21 12:59:33 x86_64 ntpd[1905]: getaddrinfo: "2.fedora.pool.ntp.org" 
invalid host address, ignored

Dec 21 13:04:34 x86_64 ntpd[1905]: Listening on interface #2 eth0, 
fe80::20f:eaff:fe8c:135c#123 Enabled
Dec 21 13:04:34 x86_64 ntpd[1905]: Listening on interface #3 eth0, 
172.16.1.38#123 Enabled

Dec 23 14:09:21 x86_64 ntpd[1905]: ntpd exiting on signal 15

ntpd[1905]: ntpd 4.2.4p4@1.1520-o Tue Apr 15 14:22:43 UTC 2008 (1)

Dec 23 14:10:41 x86_64 ntpd[1906]: precision = 1.000 usec
Dec 23 14:10:41 x86_64 ntpd[1906]: Listening on interface #0 wildcard, 
0.0.0.0#123 Disabled
Dec 23 14:10:41 x86_64 ntpd[1906]: Listening on interface #1 wildcard, 
::#123 Disabled
Dec 23 14:10:41 x86_64 ntpd[1906]: kernel time sync status 0040
Dec 23 14:10:41 x86_64 ntpd[1906]: frequency initialized 0.000 PPM from 
/var/lib/ntp/drift
Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "0.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "1.fedora.pool.ntp.org" 
invalid host address, ignored
Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "2.fedora.pool.ntp.org" 
invalid host address, ignored

ntpd[1906]: Listening on interface #2 eth0, fe80::20f:eaff:fe8c:135c#123 
Enabled
Dec 23 14:15:42 x86_64 ntpd[1906]: Listening on interface #3 eth0, 
172.16.1.38#123 Enabled

Dec 21 12:29:40 x86_64 ntpd[1891]: Deleting interface #3 eth2, 
fe80::20f:eaff:fe8c:158d#123, interface stats: received=0, sent=0, 
dropped=0, active_time=1361211 secs

[Tim]

On Tue, 2008-12-23 at 18:38 -0500, Jim wrote:
> Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "0.fedora.pool.ntp.org" invalid host address, ignored
> Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "1.fedora.pool.ntp.org" invalid host address, ignored
> Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "2.fedora.pool.ntp.org" invalid host address, ignored

Are you using NetworkManager to handle your network?  In Fedora 9 I
found that NTPD was started before the network was brought up, and NTPD
never tried again after the network came to life.  There were at least
two solutions:

1. Stop using NetworkManager (use the network service, instead).
     2. Restart NTPD after the network's alive (add a script for the
        power management daemon).

[Paul W. Frields]

On Wed, Dec 24, 2008 at 10:41:50AM +1030, Tim wrote:
> On Tue, 2008-12-23 at 18:38 -0500, Jim wrote:
> > Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "0.fedora.pool.ntp.org" invalid host address, ignored
> > Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "1.fedora.pool.ntp.org" invalid host address, ignored
> > Dec 21 12:41:34 x86_64 ntpd[1906]: getaddrinfo: "2.fedora.pool.ntp.org" invalid host address, ignored
>
> Are you using NetworkManager to handle your network?  In Fedora 9 I
> found that NTPD was started before the network was brought up, and NTPD
> never tried again after the network came to life.  There were at least
> two solutions:
>  1. Stop using NetworkManager (use the network service, instead).
>  2. Restart NTPD after the network's alive (add a script for the
>     power management daemon).

In that case, to fix this, you can probably run system-config-time
and, under the NTP tab, and Advanced Options, turn off the option
labeled "Synchronize clock when system starts."

To make sure that the clock is relatively in sync for the future
without having to take a long time for the skew to be resolved, you
can do this:

su -                                # provide root password
service ntpd stop
ntpdate 0.fedora.pool.ntp.org       # or your server of choice
service ntpd start
exit

[Todd Zullinger]

Jim wrote:
> Mikkel L. Ellertson wrote:
> > To you have a full-time internet connection?
>
> yes

If the ntpd service starts before NetworkManager brings up your
network connection, ntpd doesn't always seem to be able to connect.
If you restart ntpd after your connection is brought up, does it work
then?  (If your time is way off, you might need to sync it manually,
using "ntpdate -u $some_time_server".)

[Todd Zullinger]

Paul W. Frields wrote:
> su -                                # provide root password
> service ntpd stop
> ntpdate 0.fedora.pool.ntp.org       # or your server of choice
> service ntpd start
> exit

I presume you're stopping ntpd because otherwise ntpdate complains
that "the NTP socket is in use, exiting"?

To avoid that, just use the -u option to ntpdate and it will be able
to update while ntpd is running.

[Nifty Fedora Mitch]

On Tue, Dec 23, 2008 at 3:38 PM, Jim  wrote:
> Mikkel L. Ellertson wrote:
> > Jim wrote:
> > > Fedora 8 won't keep correct time.
> > > "Unable to connect to time server"  , it doesn' make any difference what
> > > time server you try to connect to.
> > > "ntpd" is running in services.

/////////
> Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "0.fedora.pool.ntp.org"
> invalid host address, ignored
> Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "1.fedora.pool.ntp.org"
> invalid host address, ignored
> Dec 23 14:10:41 x86_64 ntpd[1906]: getaddrinfo: "2.fedora.pool.ntp.org"
> invalid host address, ignored

This tells me that you are not connected to an effective and valid
name server at the time
you are starting NTP.
> Dec 21 12:29:40 x86_64 ntpd[1891]: Deleting interface #3 eth2,
> fe80::20f:eaff:fe8c:158d#123, interface stats: received=0, sent=0,
> dropped=0, active_time=1361211 secs

With eth2 I wonder what is dialed in with eth0 and eth1.
A system with multiple interfaces could be routing in
ways that never get connected to anything interesting.

Are you posting from this same host or
from another connection?  If you can email some
things are live.   It may be that some stuff
is not setup at boot time resulting in an error
and later when you get connected NTP has shut down for
want of good services.   If this is the case restart it now
that you are connected...

[Paul W. Frields]

On Tue, Dec 23, 2008 at 07:32:44PM -0500, Todd Zullinger wrote:
> Paul W. Frields wrote:
> > su -                                # provide root password
> > service ntpd stop
> > ntpdate 0.fedora.pool.ntp.org       # or your server of choice
> > service ntpd start
> > exit
>
> I presume you're stopping ntpd because otherwise ntpdate complains
> that "the NTP socket is in use, exiting"?
>
> To avoid that, just use the -u option to ntpdate and it will be able
> to update while ntpd is running.

Good tip, thanks Todd!