My onsite server is always 15 min slow and seems like NTPD (Network Time Protocol (NTP) daemon) not running properly. can anyone suggest me how to rectify this problem? we can't seem to get NTP to properly sync the clock.
Any help is resolving the issue will be helpful. Thanks in advance to all
First, I would manually adjust the clock to be closer. You're 900 seconds off assuming your 15 minutes is accurate. ntp maxes out at about 1000 seconds. Next, any errors in the logs? And check your ntp config: use a good server and use a drift file.
ntp gives up completely if your clock is too far off. You should set your clock so it is as close as you can get it. Is your clock 5 seconds slow or fast? No problem. Is your clock 1001 seconds(or more) too slow or too fast? Forget ntp... it will not even try to sync the clock.
ntp needs to figure out how bad your clock is. If you reboot a lot, it has to keep starting from zero. With a drift file, it writes the drift into the file. Later when ntp restarts, it rereads the drift file. So it doesn't need to start completely over.
What server are you using? Post your whole ntp config file.
This configuration looks like it is intended as the ntp server for your organization. It obtains the time from external sources. And it is will to distribute the time to other boxes. So there might be other boxes that use this box as their ntp server. Or maybe this box is the only system in your organization.
server 127.127.1.0 # local clock
fudge 127.127.1.0 stratum 10
This is a NOT great idea in your situation. You have specified, that in the special case where no other servers can be contacted, to use the local clock as if it were a server. Well, you have also stated that your local clock is off by 15 minutes. Any other boxes that get ntp through this box will try to sync to this box's local clock. If you remove those lines, if this box can't contact a server, you have no ntp service. Everyone just uses their local clocks. If this box does not serve others, there is no harm, but I still would not do it. This should be done only on systems with great internals clocks.
driftfile /etc/ntp/drift
You are using a driftfile. That is good.
server 132.249.20.60 # billthecat.sdsc.edu
server 204.152.184.72 # clock.isc.org
I did not find the first site on the lists. The second site is a stratum 1. I would just use the second entry. If you use multiple servers, you need at least 3. Two servers does not work very well.
Read the NTP FAQ, especially the part about NTP in real life.
I am talking about billthecat. When I said I could not find it on the "lists", I did not mean the lists of boxes which you can successfully ping. I meant the official lists of ntp servers.
you know thats what I fear - in search of a good document, I will endup with lot many unnecessary docs .. Thats why I asked my Linux friends if someone knows about any good docs, anyway no probs I will find it out ...
.. Thats why I asked my Linux friends if someone knows about any good docs, anyway no probs I will find it out ...