How to run the NTP daemon automatically?

Question

On my RaspBMC, the time drifts pretty much (more than a day in one month). So I started to use NTP, but have problems running it.

I called sudo update-rc.d ntp enable, I also added DAEMONS=(!hwclock ntp ntpdate) to /etc/rc.local, I made sure there is a correct symlink from /etc/rc2.d/S18ntp, however, after rebooting, service ntp status says the daemon is not running. Running the daemon manually by calling sudo service ntp start works correctly.

What else should I do to run the daemon all the time?

Edit: In /var/log/daemon.log I see normal functioning when I run the service manually, but after a reboot, this is shown:

Mar  3 08:57:46 raspbmc ntpd[1179]: ntpd [email protected] Sat Dec 20 21:33:54 UTC 2014 (1)
Mar  3 08:57:46 raspbmc ntpd[1202]: proto: precision = 1.000 usec
Mar  3 08:57:46 raspbmc ntpd[1202]: unable to bind to wildcard address 0.0.0.0 - another process may be running - EXITING
Mar  3 08:57:46 raspbmc init: ntp main process (1202) terminated with status 1
Mar  3 08:57:46 raspbmc init: ntp main process ended, respawning
...
Mar  3 08:57:47 raspbmc init: ntp respawning too fast, stopped

Answer 1

All you should need is update-rc.d So I'm guessing you or raspbmc inadvertently changed something from default debian/raspbian/rpi-raspbian making one of the tests in /etc/init.d/ntp fail.

@ppumkin said

sudo apt-get remove ntpdate

As It can conflict. (they use the same port) You can also just disable it's invocation if you want to keep using it.

This is what it looks like on my raspbmc where ntp seems to be working

root@raspbmc:~# service ntp status
ntp start/running, process 941

root@raspbmc:~# grep ntp /var/log/daemon.log | tail                            
Jan 31 11:27:24 raspbmc ntpd[920]: ntpd [email protected] Sat Dec 20 21:33:54 UTC 2014 (1)
Jan 31 11:27:24 raspbmc ntpd[941]: proto: precision = 1.000 usec
Jan 31 11:27:24 raspbmc ntpd[941]: Listen and drop on 0 v4wildcard 0.0.0.0 UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: Listen and drop on 1 v6wildcard :: UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: Listen normally on 2 lo 127.0.0.1 UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: Listen normally on 3 eth0 192.168.0.153 UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: Listen normally on 4 lo ::1 UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: Listen normally on 5 eth0 fe80::ba27:ebff:fec1:c508 UDP 123
Jan 31 11:27:24 raspbmc ntpd[941]: peers refreshed
Jan 31 11:27:24 raspbmc ntpd[941]: Listening on routing socket on fd #22 for interface updates

root@raspbmc:~# dpkg -l | grep -i ntp
ii  ntp                             1:4.2.6.p5+dfsg-2+deb7u1           armhf        Network Time Protocol daemon and utility programs

root@raspbmc:~# find /etc/ -iname '*ntp*'
/etc/dhcp/dhclient-exit-hooks.d/ntp
/etc/init/ntp.conf
/etc/rc2.d/S17ntp
/etc/rc5.d/S17ntp
/etc/rc4.d/S17ntp
/etc/cron.daily/ntp
/etc/default/ntp
/etc/ntp.conf
/etc/rc3.d/S17ntp
/etc/init.d/ntp

root@raspbmc:~# nice grep -RiP "[^a-z]ntp" /etc/ 2>/dev/null | perl -pe 's/:.*//g'  | grep -v ntp | sort -u 
/etc/dhcp/dhclient-enter-hooks.d/debug
/etc/dhcp/dhclient-exit-hooks.d/debug
/etc/dhcp/dhclient.conf
/etc/init.d/.depend.start
/etc/init/early-network-time.conf
/etc/passwd
/etc/passwd-
Binary file /etc/alternatives/net matches

root@raspbmc:~# runlevel
 2

Start by looking at the log and comparing the other commands I showed, then if nothing jumps out add /bin/date >> /home/pi/ntp.log on the 11th line of /etc/init.d/ntp to see if the service is being called at all, if not, remove soft links and re-run update-rc.d if it is then move the log line down the script file until you find where it's failing. Use a binary like search for the log line position (0%,50%,100%,75%,62%,etc)

Answer 2

It seems like the application ntpupdate might be running during that rc level.

Please try removing it and reboot.

sudo apt-get remove ntpdate

NTP is its successor but it seems ntpupdate to be installed and run by default.