ntpd question

jim dorey <skaar.DeleteThis@ns.sympatico.ca> wrote:
> root wrote:
>> How long does it take for the daemon to fix the
>> clock drift? The daemon has been running for
>> 9 hours and my clock is off by 6 minutes. Without
>> ntpd my clock would have been off by roughly the
>> same amount.
>
> if you gotta, run ntpdate before starting ntpd...i've got ntpdate in a
> cron job set for daily, i get a drift of a few seconds a day, i'm not
> hyper critical on that, like i used to be.

I did run ntpdate just before I started ntpd. Nine hours later
the entry in /etc/ntp/drift was still 0.000. When does it get
changed?

root <NoEMail.TakeThisOut@home.org> wrote:
> How long does it take for the daemon to fix the
> clock drift? The daemon has been running for
> 9 hours and my clock is off by 6 minutes. Without
> ntpd my clock would have been off by roughly the
> same amount.

If so it seems as if ntpd hasn't succeeded in keeping your clock
synchronized. What output do you get from ntptrace?

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc3(at)poolhem.se Examples of addresses which go to spammers:
root@localhost postmaster@localhost

Henrik Carlqvist <Henrik.Carlqvist.RemoveThis@deadspam.com> wrote:
> root <NoEMail.RemoveThis@home.org> wrote:
>> How long does it take for the daemon to fix the
>> clock drift? The daemon has been running for
>> 9 hours and my clock is off by 6 minutes. Without
>> ntpd my clock would have been off by roughly the
>> same amount.
>
> If so it seems as if ntpd hasn't succeeded in keeping your clock
> synchronized. What output do you get from ntptrace?
>
> regards Henrik

I just re-started ntpd, after deciding it wasn't helping.
After a few seconds of operation ntptrace yields:
localhost: stratum 16, offset 0.000000, synch distance 0.000210

I don't have a man entry for ntptrace, what should I be looking
for?

Henrik Carlqvist <Henrik.Carlqvist.RemoveThis@deadspam.com> wrote:
> root <NoEMail.RemoveThis@home.org> wrote:
>> How long does it take for the daemon to fix the
>> clock drift? The daemon has been running for
>> 9 hours and my clock is off by 6 minutes. Without
>> ntpd my clock would have been off by roughly the
>> same amount.
>
> If so it seems as if ntpd hasn't succeeded in keeping your clock
> synchronized. What output do you get from ntptrace?
>
> regards Henrik

I read the man page for ntptrace and for:
ntptrace 0.pool.ntp.org
jabber.sotelips.net: timed out, nothing received
***Request timed out

Similarly for:
ntptrace tick.ucla.edu
tick.ucla.edu: timed out, nothing received
***Request timed out

I can ping both of these servers and they are alive.
Moreover, I ran ntpdate to tick.ucla.edu just
before starting ntpd. The daemon opens and locks
a socket, but it isn't talking to anyone.

On 11/02/09 11:44, root wrote:

> I just re-started ntpd, after deciding it wasn't helping. After a
> few seconds of operation ntptrace yields:
> localhost: stratum 16, offset 0.000000, synch distance 0.000210
>
> I don't have a man entry for ntptrace, what should I be looking
> for?

It simply means that your ntp server can't get to any time server and
it it is synchronizing on the local clock.

Is your network working?

Ciao
Giovanni
--
A computer is like an air conditioner,
it stops working when you open Windows.
< http://giovanni.homelinux.net/ >

root <NoEMail RemoveThis @home.org> wrote:
> Henrik Carlqvist <Henrik.Carlqvist RemoveThis @deadspam.com> wrote:
>> root <NoEMail RemoveThis @home.org> wrote:
>>> How long does it take for the daemon to fix the
>>> clock drift? The daemon has been running for
>>> 9 hours and my clock is off by 6 minutes. Without
>>> ntpd my clock would have been off by roughly the
>>> same amount.
>>
>> If so it seems as if ntpd hasn't succeeded in keeping your clock
>> synchronized. What output do you get from ntptrace?
>>
>> regards Henrik
>
> I read the man page for ntptrace and for:
> ntptrace 0.pool.ntp.org
> jabber.sotelips.net: timed out, nothing received
> ***Request timed out
>
> Similarly for:
> ntptrace tick.ucla.edu
> tick.ucla.edu: timed out, nothing received
> ***Request timed out
>
> I can ping both of these servers and they are alive.
> Moreover, I ran ntpdate to tick.ucla.edu just
> before starting ntpd. The daemon opens and locks
> a socket, but it isn't talking to anyone.

I had started ntpd with an old /etc/rc.d/rc.ntpd script.
When I use rc.ntpd.new I can get a connection. Here is
what I see:

ntptrace tick.ucla.edu
tick.ucla.edu: timed out, nothing received
***Request timed out

ntptrace 0.pool.ntp.org
0.pool.ntp.org: stratum 2, offset -0.000071, synch distance 0.045360
clock.trit.net: timed out, nothing received
***Request timed out

ntptrace 1.pool.ntp.org
1.pool.ntp.org: timed out, nothing received
***Request timed out

ntptrace 2.pool.ntp.org
2.pool.ntp.org: stratum 2, offset 0.000666, synch distance 0.073360
***Association ID 54222 unknown to server

ntptrace 3.pool.ntp.org
3.pool.ntp.org: stratum 3, offset 0.000119, synch distance 0.044794
10.1.11.61: timed out, nothing received
***Request timed out

Without any argument to ntptrace I just get:
localhost: stratum 11, offset 0.000000, synch distance 0.206585

At this time my /etc/ntp.conf file is:
server 127.127.1.0 # local clock
fudge 127.127.1.0 stratum 10
multicastclient # listen on default 224.0.1.1
broadcastdelay 0.008
keys /etc/ntp/keys
trustedkey 65535
requestkey 65535
controlkey 65535
driftfile /etc/ntp/drift
pidfile /var/run/ntpd.pid
server 0.pool.ntp.org
server 1.pool.ntp.org
server 2.pool.ntp.org
server 3.pool.ntp.org
server tick.ucla.edu

On 11/02/09 12:23, root wrote:
> root <NoEMail.DeleteThis@home.org> wrote:
>> Henrik Carlqvist <Henrik.Carlqvist.DeleteThis@deadspam.com> wrote:
>>> root <NoEMail.DeleteThis@home.org> wrote:
>>>> How long does it take for the daemon to fix the
>>>> clock drift? The daemon has been running for
>>>> 9 hours and my clock is off by 6 minutes. Without
>>>> ntpd my clock would have been off by roughly the
>>>> same amount.
>>> If so it seems as if ntpd hasn't succeeded in keeping your clock
>>> synchronized. What output do you get from ntptrace?
>>>
>>> regards Henrik
>> I read the man page for ntptrace and for:
>> ntptrace 0.pool.ntp.org
>> jabber.sotelips.net: timed out, nothing received
>> ***Request timed out
>>
>> Similarly for:
>> ntptrace tick.ucla.edu
>> tick.ucla.edu: timed out, nothing received
>> ***Request timed out
>>
>> I can ping both of these servers and they are alive.
>> Moreover, I ran ntpdate to tick.ucla.edu just
>> before starting ntpd. The daemon opens and locks
>> a socket, but it isn't talking to anyone.
>
> I had started ntpd with an old /etc/rc.d/rc.ntpd script.
> When I use rc.ntpd.new I can get a connection. Here is
> what I see:
>
> ntptrace tick.ucla.edu
> tick.ucla.edu: timed out, nothing received
> ***Request timed out
>
> ntptrace 0.pool.ntp.org
> 0.pool.ntp.org: stratum 2, offset -0.000071, synch distance 0.045360
> clock.trit.net: timed out, nothing received
> ***Request timed out
>
> ntptrace 1.pool.ntp.org
> 1.pool.ntp.org: timed out, nothing received
> ***Request timed out
>
> ntptrace 2.pool.ntp.org
> 2.pool.ntp.org: stratum 2, offset 0.000666, synch distance 0.073360
> ***Association ID 54222 unknown to server
>
> ntptrace 3.pool.ntp.org
> 3.pool.ntp.org: stratum 3, offset 0.000119, synch distance 0.044794
> 10.1.11.61: timed out, nothing received
> ***Request timed out
>
> Without any argument to ntptrace I just get:
> localhost: stratum 11, offset 0.000000, synch distance 0.206585
>
> At this time my /etc/ntp.conf file is:
> server 127.127.1.0 # local clock
> fudge 127.127.1.0 stratum 10
> multicastclient # listen on default 224.0.1.1
> broadcastdelay 0.008
> keys /etc/ntp/keys
> trustedkey 65535
> requestkey 65535
> controlkey 65535
> driftfile /etc/ntp/drift
> pidfile /var/run/ntpd.pid
> server 0.pool.ntp.org
> server 1.pool.ntp.org
> server 2.pool.ntp.org
> server 3.pool.ntp.org
> server tick.ucla.edu

My ntp.conf file looks as follows:
## Remote servers (I'm in northern Italy )
server it.pool.ntp.org ## Stratum 2/3 servers
server ch.pool.ntp.org ## Stratum 2/3 servers
## Use local server from LAN as preferred
server milena.home.net.it prefer

## Local clock source
server 127.127.1.0
fudge 127.127.1.0 stratum 10

## Drift file
driftfile /etc/ntp/drift

#multicastclient # listen on default 224.0.1.1
#broadcastdelay 0.008

## Keys file
keys /etc/ntp/ntp.keys
#trustedkey 65535
#requestkey 65535
#controlkey 65535

## Don't serve time or stats or trust anyone else
restrict default noquery nomodify # notrust
restrict localhost

## Work files
# logfile /var/log/ntpd.log
pidfile /var/run/ntpd.pid

## end ntp.conf


I'm not sure, but I think you have to comment the keys ID and the
multicast ditective. ntptrace should give the hierarchy of used ntp
servers.

# root@monica:~ > ntptrace
localhost: stratum 3, offset -0.001629, synch distance 0.073363
milena.home.net.it: stratum 2, offset -0.000408, synch \
distance 0.035561
ntp1.inrim.it: stratum 1, offset -0.000033, synch \
distance 0.000138, refid 'UTCI'
# root@monica:~ >

Ciao
Giovanni
--
A computer is like an air conditioner,
it stops working when you open Windows.
< http://giovanni.homelinux.net/ >

Giovanni <lsodgf0.DeleteThis@home.net.it> wrote:
>
> I'm not sure, but I think you have to comment the keys ID and the
> multicast ditective. ntptrace should give the hierarchy of used ntp
> servers.
>
> # root@monica:~ > ntptrace
> localhost: stratum 3, offset -0.001629, synch distance 0.073363
> milena.home.net.it: stratum 2, offset -0.000408, synch \
> distance 0.035561
> ntp1.inrim.it: stratum 1, offset -0.000033, synch \
> distance 0.000138, refid 'UTCI'
> # root@monica:~ >
>
> Ciao
> Giovanni

I have tried commenting out the keys stuff, but the
ntptrace command, without specifying a server, never
goes anywhere.

I have been running a sequence of these steps:
source /etc/rc.d/rc.ntpd stop
modify /etc/ntp.conf
source /etc/rc.d/rc.ntpd start
ntptrace

Repeat over and over. So far no luck.

I'll try your conf file as is.

root <NoEMail.TakeThisOut@home.org> wrote:
> Giovanni <lsodgf0.TakeThisOut@home.net.it> wrote:
>>
>> I'm not sure, but I think you have to comment the keys ID and the
>> multicast ditective. ntptrace should give the hierarchy of used ntp
>> servers.
>>
>> # root@monica:~ > ntptrace
>> localhost: stratum 3, offset -0.001629, synch distance 0.073363
>> milena.home.net.it: stratum 2, offset -0.000408, synch \
>> distance 0.035561
>> ntp1.inrim.it: stratum 1, offset -0.000033, synch \
>> distance 0.000138, refid 'UTCI'
>> # root@monica:~ >
>>
>> Ciao
>> Giovanni
>
> I have tried commenting out the keys stuff, but the
> ntptrace command, without specifying a server, never
> goes anywhere.
>
> I have been running a sequence of these steps:
> source /etc/rc.d/rc.ntpd stop
> modify /etc/ntp.conf
> source /etc/rc.d/rc.ntpd start
> ntptrace
>
> Repeat over and over. So far no luck.
>
> I'll try your conf file as is.

OK. No change with your config file, as is.
I have no firewall stuff to block the port,
and I think the fact that ntpdate can talk
to the servers that communication is possible.

root wrote:
> How long does it take for the daemon to fix the
> clock drift? The daemon has been running for
> 9 hours and my clock is off by 6 minutes. Without
> ntpd my clock would have been off by roughly the
> same amount.
>
Try this: in //etc/ntp.conf/ change the servers section to look like
this (assuming you're in the US, otherwise, change the country code as
appropriate)

server 127.127.1.0 # local clock
fudge 127.127.1.0 stratum 10
#server pool.ntp.org
server 0.us.pool.ntp.org
server 1.us.pool.ntp.org
server 2.us.pool.ntp.org

(leave the local clock and fudge alone, they're there for local sync
when the network goes away for whatever reason)

What we're doing is giving the daemon three external time servers to
"vote" for the "best" time source.

and edit //etc/rc.d/rc.ntpd/ adding

# Start ntpd:
ntpd_start() {
*# We add a log for NTP; it gets wiped out on every daemon start*
* >/tmp/ntp.log *
CMDLINE="/usr/sbin/ntpd -g"
echo -n "Starting NTP daemon: $CMDLINE"
$CMDLINE -p /var/run/ntpd.pid *-l /tmp/ntp.log *
echo
}

(this gives you log entries in //tmp/ntp.log/)

Then, stop and restart NTP (//etc/rc.d/rc.ntpd stop/, wait a few
seconds, //etc/rc.d/rc.ntpd start/)

After a minute or so your //tmp/ntp.log/ will look similar to this

cat /tmp/ntp.log
2 Nov 08:18:03 ntpd[3007]: logging to file /tmp/ntp.log
2 Nov 08:18:03 ntpd[3007]: precision = 1.000 usec
2 Nov 08:18:03 ntpd[3007]: ntp_io: estimated max descriptors: 1024,
initial socket boundary: 16
2 Nov 08:18:03 ntpd[3007]: Listening on interface #0 wildcard,
0.0.0.0#123 Disabled
2 Nov 08:18:03 ntpd[3007]: Listening on interface #1 lo,
127.0.0.1#123 Enabled
2 Nov 08:18:03 ntpd[3007]: Listening on interface #2 eth0,
192.168.1.30#123 Enabled
2 Nov 08:18:03 ntpd[3007]: Listening on interface #3 ppp0,
69.54.207.108#123 Enabled
2 Nov 08:18:03 ntpd[3007]: kernel time sync status 0040
2 Nov 08:18:03 ntpd[3007]: frequency initialized 0.000 PPM from
/etc/ntp/drift
2 Nov 08:21:18 ntpd[3007]: synchronized to LOCAL(0), stratum 10
2 Nov 08:21:18 ntpd[3007]: kernel time sync status change 0001
2 Nov 08:22:24 ntpd[3007]: synchronized to 65.255.217.202, stratum 3


After a few minutes, you ought to see something similar to this by
executing /nptq -p/

remote refid st t when poll reach delay
offset jitter
==============================================================================
LOCAL(0) .LOCL. 10 l 34 64 377 0.000
0.000 0.001
+www.broadbandja 69.63.177.217 3 u 34 64 377 236.961
16.837 29.095
+router-gw1.net. 192.43.244.18 2 u 34 64 377 286.630
66.475 34.978
*198.186.191.229 64.147.116.229 2 u 32 64 377 220.962
22.407 12.225

The asterisk indicates the server you're synchronized to, the plus signs
indicate candidates. By defining three external servers (in
//etc/ntp.conf/), we're giving the daemon a range of servers so it can
select the best one for synchronization (least delay time, least
jitter). Over time, the servers shown by /ntpq -p/ will change as other,
"better" servers become available (the log will show when this happens).

If you don't get synchronized in, oh, fifteen minutes or so you clock
may be too far off for NTP to adjust it; NTP won't slew a clock six
minutes for example.

So, stop the daemon, execute /ntpdate/ using one of the pool servers you
see from /ntpq -p/ then start the daemon again. It may take some time
(hours) for //etc/ntp/drift/ to change in any event.

Hope this helps some.

Thomas Ronayne wrote:
> root wrote:
>> How long does it take for the daemon to fix the
>> clock drift? The daemon has been running for
>> 9 hours and my clock is off by 6 minutes. Without
>> ntpd my clock would have been off by roughly the
>> same amount.
>>
>
> # Start ntpd:
> ntpd_start() {
> *# We add a log for NTP; it gets wiped out on every daemon start*
> * >/tmp/ntp.log *
> CMDLINE="/usr/sbin/ntpd -g"
> echo -n "Starting NTP daemon: $CMDLINE"
> $CMDLINE -p /var/run/ntpd.pid *-l /tmp/ntp.log *
> echo
> }
> (this gives you log entries in //tmp/ntp.log/)
>
>
Well, that got pretty messy... should look like this
# Start ntpd:
ntpd_start() {
# We add a log for NTP; it gets wiped out on every daemon start
>/tmp/ntp.log
CMDLINE="/usr/sbin/ntpd -g"
echo -n "Starting NTP daemon: $CMDLINE"
$CMDLINE -p /var/run/ntpd.pid -l /tmp/ntp.log
echo
}

Thomas Ronayne <trona RemoveThis @ameritech.net> wrote:
> Hope this helps some.

I think it was a big help. I get this in the log file
after re-starting ntpd:
2 Nov 05:51:24 ntpd[3077]: logging to file /tmp/ntp.log
2 Nov 05:51:24 ntpd[3077]: precision = 4000.000 usec
2 Nov 05:51:24 ntpd[3077]: ntp_io: estimated max descriptors: 1024, initial socket boundary: 16
2 Nov 05:51:24 ntpd[3077]: Listening on interface #0 wildcard, 0.0.0.0#123 Disabled
2 Nov 05:51:24 ntpd[3077]: Listening on interface #1 lo, 127.0.0.1#123 Enabled
2 Nov 05:51:24 ntpd[3077]: Listening on interface #2 eth0, 10.0.0.3#123 Enabled
2 Nov 05:51:24 ntpd[3077]: kernel time sync status 0040
2 Nov 05:51:25 ntpd[3077]: frequency initialized 1.211 PPM from /etc/ntp/drift
2 Nov 05:54:41 ntpd[3077]: synchronized to LOCAL(0), stratum 10
2 Nov 05:54:41 ntpd[3077]: kernel time sync status change 0001

and I get this for ntpq -p:
remote refid st t when poll reach delay offset jitter
==============================================================================
*LOCAL(0) .LOCL. 10 l 46 64 77 0.000 0.000 3.906
clock.team-cymr 192.168.0.16 2 u 47 64 77 67.367 27839.7 2012.35
lime7.adamantsy 208.75.88.4 3 u 41 64 77 48.759 27902.8 2043.38
rrcs-64-183-56- .GPS. 1 u 50 64 77 23.623 27810.4 1974.96

So it seems that it may be seeing 3 servers. I'll let it run for
a few hours to see if the drift entry changes.

Thanks a whole lot.

Thomas Ronayne <trona.DeleteThis@ameritech.net> wrote:
> Thomas Ronayne wrote:
>> root wrote:
>>> How long does it take for the daemon to fix the
>>> clock drift? The daemon has been running for
>>> 9 hours and my clock is off by 6 minutes. Without
>>> ntpd my clock would have been off by roughly the
>>> same amount.
>>>
>>
>> # Start ntpd:
>> ntpd_start() {
>> *# We add a log for NTP; it gets wiped out on every daemon start*
>> * >/tmp/ntp.log *
>> CMDLINE="/usr/sbin/ntpd -g"
>> echo -n "Starting NTP daemon: $CMDLINE"
>> $CMDLINE -p /var/run/ntpd.pid *-l /tmp/ntp.log *
>> echo
>> }
>> (this gives you log entries in //tmp/ntp.log/)
>>
>>
> Well, that got pretty messy... should look like this
> # Start ntpd:
> ntpd_start() {
> # We add a log for NTP; it gets wiped out on every daemon start
> >/tmp/ntp.log
> CMDLINE="/usr/sbin/ntpd -g"
> echo -n "Starting NTP daemon: $CMDLINE"
> $CMDLINE -p /var/run/ntpd.pid -l /tmp/ntp.log
> echo
> }
>

Thanks again, I got it from your first post.