My ntp.conf file gets overwritten!

Peter

unread,

Dec 22, 2003, 4:39:16 PM12/22/03

to

I have a simple two line ntp.conf file

server tick.jrc.us
driftfile /etc/ntp/ntp.drift

ntp works fine and sets time. At some point, an entirely new ntp.conf file
is written to /etc and of course, time stops updating. I grepped for the
file but could not find it.

Here is the replaced ntp,conf

*** ntp.conf replaced ***
restrict default noquery notrust nomodify
restrict 127.0.0.1
restrict 192.168.2.0 mask 255.255.255.0
fudge 127.127.1.0 stratum 3
server 127.127.1.0
driftfile /etc/ntp.drift
logfile /var/log/ntp.log
*** end file ***

Where can I look to see why this is occurring?

Linux version slackware 9.1, kernel 2.4.23, ntpd: ntpd 4....@1.892 Thu Sep
11 23:47:04 PDT 2003 (1)

Here is today's ntp.log. Somewhere along the line, ntp.conf was replaced.

*** ntp.log ***
22 Dec 05:55:10 ntpd[2821]: signal_no_reset: signal 17 had flags 4000000
22 Dec 05:58:25 ntpd[2820]: time set -17.143459 s
22 Dec 05:58:25 ntpd[2820]: synchronisation lost
22 Dec 06:00:17 ntpd[2820]: ntpd exiting on signal 15
22 Dec 06:00:42 ntpd[2907]: signal_no_reset: signal 17 had flags 4000000
22 Dec 06:01:40 ntpd[2906]: ntpd exiting on signal 2
22 Dec 06:07:30 ntpd[3010]: signal_no_reset: signal 17 had flags 4000000
22 Dec 06:11:02 ntpd[3009]: time set -0.037555 s
22 Dec 06:11:02 ntpd[3009]: synchronisation lost
22 Dec 06:26:21 ntpd[3009]: kernel time discipline status change 1
22 Dec 08:58:18 ntpd[3009]: ntpd exiting on signal 15
22 Dec 09:02:22 ntpd[1120]: ntpd exiting on signal 15
22 Dec 09:08:31 ntpd[1120]: time set 0.000000 s
22 Dec 09:08:31 ntpd[1120]: synchronisation lost
22 Dec 09:18:18 ntpd[1120]: ntpd exiting on signal 15
22 Dec 10:19:40 ntpd[1120]: time set 0.000000 s
22 Dec 10:19:40 ntpd[1120]: synchronisation lost
22 Dec 10:55:21 ntpd[1120]: ntpd exiting on signal 15
22 Dec 11:00:29 ntpd[1120]: time set 0.000000 s
22 Dec 11:00:29 ntpd[1120]: synchronisation lost
22 Dec 12:45:58 ntpd[1120]: ntpd exiting on signal 15
22 Dec 12:50:26 ntpd[1120]: kernel time discipline status change 41

Thanks and Happy Holidays. Pete

Wolfgang S. Rupprecht

unread,

Dec 22, 2003, 4:55:42 PM12/22/03

to

Peter <nos...@dontreply.com> writes:
> *** ntp.conf replaced ***
> restrict default noquery notrust nomodify
> restrict 127.0.0.1
> restrict 192.168.2.0 mask 255.255.255.0
> fudge 127.127.1.0 stratum 3
> server 127.127.1.0
> driftfile /etc/ntp.drift
> logfile /var/log/ntp.log
> *** end file ***

You might want to add an un-restriction for your local clocks:

restrict 127.127.0.0 mask 255.255.0.0 # internal clocks

-wolfgang
--
Wolfgang S. Rupprecht http://www.wsrcc.com/wolfgang/
The above "From:" address is valid. Don't mess with it.

David

unread,

Dec 22, 2003, 5:14:29 PM12/22/03

to

Peter wrote:
> I have a simple two line ntp.conf file
>
> server tick.jrc.us
> driftfile /etc/ntp/ntp.drift
>
> ntp works fine and sets time. At some point, an entirely new ntp.conf file
> is written to /etc and of course, time stops updating. I grepped for the
> file but could not find it.
>
> Here is the replaced ntp,conf
>
> *** ntp.conf replaced ***
> restrict default noquery notrust nomodify
> restrict 127.0.0.1
> restrict 192.168.2.0 mask 255.255.255.0
> fudge 127.127.1.0 stratum 3
> server 127.127.1.0
> driftfile /etc/ntp.drift
> logfile /var/log/ntp.log
> *** end file ***
>
> Where can I look to see why this is occurring?

man dhcpcd
-N Prevents dhcpcd from replacing existing /etc/ntp.conf file.

Add the "-N" option to the dhcpcd line in /etc/rc.d/rc.inet1

/sbin/dhcpcd -t -N 10 -h ${DHCP_HOSTNAME[$1]} -d eth${1}
else
/sbin/dhcpcd -t -N 10 -d eth${1}

--
Confucius: He who play in root, eventually kill tree.
Registered with The Linux Counter. http://counter.li.org/
Slackware 9.1.0 Kernel 2.4.23 SMP i686 (GCC) 3.3.2
Uptime: 4 days, 4:25, 2 users, load average: 1.34, 1.23, 1.16

Ryan Moore

unread,

Dec 22, 2003, 5:27:36 PM12/22/03

to

One possibility is a DHCP client daemon. Not sure about the DHCP client
on Slackware, but often you can pass "-N" to keep it from overwritting
your NTP configuration.

Read the man page on your DHCP client daemon.

/dev/rob0

unread,

Dec 22, 2003, 5:14:52 PM12/22/03

to

In article <u4-dnbUCL6U...@comcast.com>, Peter wrote:
> Where can I look to see why this is occurring?

dhcpcd (DHCP client) is doing it. You can change your dhcpcd command in
/etc/rc.d/rc.inet1 to add the -N option ("man dhcpcd"). Normally I do
not recommend editing these scripts other than rc.local, but in this
case it is necessary.
--
/dev/rob0 - preferred_email=i$((28*28+28))@softhome.net
or put "not-spam" or "/dev/rob0" in Subject header to reply

Allen McIntosh

unread,

Dec 22, 2003, 8:37:50 PM12/22/03

to

In article <Pine.LNX.4.56.0312221425350.28434@jaipur>,

Ryan Moore <rmo...@rmoore.dyndns.org> wrote:
>One possibility is a DHCP client daemon. Not sure about the DHCP client
>on Slackware, but often you can pass "-N" to keep it from overwritting
>your NTP configuration.
>Read the man page on your DHCP client daemon.

Usually it renames the file, so the original will still be around,
named ntp.conf.[something].

Peter

unread,

Dec 22, 2003, 8:42:25 PM12/22/03

to

Thanks all for the quick response. I would not have been able to debug that
so quickly. Interesting that in my Mandrake setup, this NEVER occurred!
Live and learn.