ntpd kernel reports TIME_ERROR: 0x2041: Clock Unsynchronized

bagas

bagas

Hello.
My system FreeBSD 11.4-RELEASE-p9 amd64.
I see an incomprehensible error in the log.
What is this error and how can I fix it?
The system's time is synchronized.
tail /var/log/ntp.log
23 May 16:44:06 ntpd[30355]: 80.249.145.122 local addr 1.2.2.8 -> <null>
23 May 16:44:06 ntpd[30355]: 79.120.30.43 local addr 1.2.2.8 -> <null>
23 May 16:44:07 ntpd[30464]: Listen normally on 0 igb0 1.2.2.8:123
23 May 16:44:07 ntpd[30464]: Listen normally on 1 lo0 127.0.0.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 2 lo1 127.0.1.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 3 lo1 127.0.2.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 4 lo1 127.0.3.1:123
23 May 16:44:07 ntpd[30464]: Listening on routing socket on fd #25 for interface updates
23 May 16:44:07 ntpd[30464]: kernel reports TIME_ERROR: 0x2041: Clock Unsynchronized
23 May 16:44:07 ntpd[30464]: kernel reports TIME_ERROR: 0x2041: Clock Unsynchronized
Click to expand...

cat /etc/ntp.conf
interface ignore wildcard
interface listen 1.2.2.8
disable monitor
server 0.ru.pool.ntp.org iburst maxpoll 9
server 1.ru.pool.ntp.org iburst maxpoll 9
server 2.ru.pool.ntp.org iburst maxpoll 9
logfile /var/log/ntp.log
restrict default ignore
restrict 127.0.0.1
restrict 0.ru.pool.ntp.org
restrict 1.ru.pool.ntp.org
restrict 2.ru.pool.ntp.org
Click to expand...

ntptime
ntp_gettime() returns code 0 (OK)
time e454dcab.e06b15f8 Sun, May 23 2021 16:52:43.876, (.876634814),
maximum error 131705 us, estimated error 291 us, TAI offset 0
ntp_adjtime() returns code 0 (OK)
modes 0x0 (),
offset 497.816 us, frequency 42.809 ppm, interval 1 s,
maximum error 131705 us, estimated error 291 us,
status 0x2001 (PLL,NANO),
time constant 6, precision 0.001 us, tolerance 496 ppm,
Click to expand...

ntpq -p
remote refid st t when poll reach delay offset jitter
==============================================================================
+91.206.16.3 (tm 40.52.66.2 2 u 39 64 37 78.166 -0.404 3.017
*ftpshare1.corbi 89.109.251.24 2 u 41 64 37 48.183 +0.979 0.144
ns.aksinet.net .INIT. 16 u - 64 0 0.000 +0.000 0.000
Click to expand...
 
When NTP starts, the clock by definition is unsynchronized. It can take a little while for it to synch with a source.
Based on what you've included for ntp.log, it looks like ntpd was started around 23 May 16:44:06 or maybe a little before that and the error logs are 23 May 16:44:07. That's only a second or two, so the clock has not synchronized yet. It may take minutes to synch.

If ntpq -p is showing a peer and sync, check ntp.log to see if the error logs have stopped (they should have).
I generally discount "unsynchronized" logs in the first 5 minutes of starting ntp as long as it eventually synchs.

My opinion, others may have a different one.
 
cat /etc/ntp.conf
interface listen 1.2.2.8
The service weighs on an external ip address (1.2.2.8), how to remove local ones?
How to remove support?
23 May 16:44:07 ntpd[30464]: Listen normally on 1 lo0 127.0.0.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 2 lo1 127.0.1.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 3 lo1 127.0.2.1:123
23 May 16:44:07 ntpd[30464]: Listen normally on 4 lo1 127.0.3.1:123
Click to expand...
 
Jose said:
Are you trying to set up a public NTP server? Please post your ntp.conf.
Click to expand...
His second code block has ntp.conf.

If all he needs is a client, I like using openntpd from ports/packages. A bit simpler configuration and pretty clear when it's doing client only.

By default base ntpd listens on all interfaces on the machine. "man ntp.conf" should help I think you want to look at "interface ignore" directives (which you have).
Maybe comment out the "restrict localhost" will disable the listen ports on lo
 
Jose said:
Are you trying to set up a public NTP server? Please post your ntp.conf.
Click to expand...
I only need an ntp client.
Set the correct time on the server.


cat /etc/ntp.conf
interface ignore wildcard
interface listen 1.2.2.8
disable monitor
server 0.ru.pool.ntp.org iburst maxpoll 9
server 1.ru.pool.ntp.org iburst maxpoll 9
server 2.ru.pool.ntp.org iburst maxpoll 9
logfile /var/log/ntp.log
restrict default ignore
restrict 127.0.0.1
restrict 0.ru.pool.ntp.org
restrict 1.ru.pool.ntp.org
restrict 2.ru.pool.ntp.org
 
mer said:
By default base ntpd listens on all interfaces on the machine. "man ntp.conf" should help I think you want to look at "interface ignore" directives (which you have).
Click to expand...
Did not find such a parameter interface ignore.
 
bagas said:
Did not find such a parameter interface ignore.
Click to expand...
Your config has "interface ignore wildcard" that should prevent starting a listening socket on all interfaces. Then your interface listen 1.2.2.8 should listen on 1.2.2.8. I was wondering if the "restrict 127.0.0.1" is somehow creating the listening sockets on localhost.

If you only need an ntp client so you sync with other servers, take a look at openntpd port. It's configuration is simpler.

Here's what I did:
as root
service ntpd stop
pkg install openntpd
edit the /usr/local/etc/ntp.conf (this is the config file for openntpd) to point to different pool if you want
service openntpd start
To check for sync, this may take a minute or two:
ntpctl -s status

Then make sure you do as root:
sysrc ntpd_enable="NO"
sysrc openntpd_enable="YES"
 
If the time difference it too big the clocks never synchronise.
You can force it by running ntpdate at start.
in /etc/rc.local
Code: 
ntpdate 0.be.pool.ntp.org
 
You must log in or register to reply here.
Back
Top