On Monday 04 July 2005 08:38 pm, Peter B Van Campen wrote:
Thanks for the reply. My Kernel is the current YAST provided update 2.6.11.4-21.7 and the same in the SMP flavor. The situation here is centered about the SUSE changes in the way service XNTPD handles time sync for the system. The YAST runlevel services editor is set to start xntpd to sync time upon boot, and this is observed being done. Also, I can observe the clock being updated when I use YAST to 'disable and re-enable' the service. But without manual intervention the clock is left to drift until the next boot. This may be fine for infrequently used systems that are booted each time the user wants to use them, but hell on those of us that leave them on all the time.
Hi Michael,
There is no VM involved on my systems. "ntpq -p" returns "ntpq: read: Connection refused" when the service is disabled and "No association ID's returned" when XNTPD is enabled. "ntptrace" returns "localhost: stratum 16, offset 0.000000, synch distance 0.002145" when XNTPD is enabled and "/usr/sbin/ntpq: read: Connection refused" when disabled.
Sid, could you share your cron job to re-sync every minute. Or, is there a way to get the system to regularly re-sync?
Thanks, PeterB
Running xntpd should re-sync periodically. Have never had a problem with it here. And certainly, syncing your clock every minute is very extreme. Try deleting the file: /etc/adjtime. This maybe causing your clock to drift since it sets up an amount of drift to correct when it sets the time. If if you clock was originally a ways off, it may have an overly large drift adjustment.