SuSE NTP Server Sync Issues

Background

Here's the skinny of it. Under SLES you setup an NTP Server by making changes to the NTP client settings under YaST. If you have entered in the ntp server address, tested it, and have it running as a daemon you are all set - except for one small thing. Your server will like start running out of sync.

Problem

The reason, from what I have read, is that NTP always requires two sources to synchronize time. If you only add one NTP server, as I did, the secondary clock is localtime (your hardware clock) which is obviously not a reliable source for an NTP server. When the daemon is running it will initially synchronize with the server and then periodically after that. If the server is unavailable then NTP will use the secondary source (in my case, and probably yours - the hardware clock).

Solution

If you don't want to add another secondary source for time to replace the hardware clock you can add a couple of options. The options are "burst" and "iburst". Burst sends a "burst" of eight packets to the remote server if the server is reachable (as opposed to 1 packet without the burst option). The second option, iburst, will do the same as burst except it does it when the remote server is unreachable.

To add these options simply open the YaST NTP Client module and select Complex Configuration. From there you can add burst and iburst to the options field (space delimited).

ie:

burst iburst

You can also enter in the values directly in to /etc/ntp.conf directly

server xxx.xxx.xxx.xxx burst iburst.

That's all!