[ale] Time issue

[Ryan Fish]

I really do not want to rebuild a mail server from scratch.  It is a HP Netserver LPr though...

I setup ntpd but the time is already off by almost two hours!

/var/log/messages now shows me the following:

Jan 12 16:16:10 <server> ntpd[4584]: synchronisation lost
Jan 12 17:12:50 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 18:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 19:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 20:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 21:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 22:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 12 23:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied
Jan 13 00:12:51 <server> ntpd[4584]: can't open /etc/ntp/drift.TEMP: Permission denied

I stopped ntpd and reset the system time using 'ntpdate'.  I then restarted ntpd.

/var/log/messages now has this:

Jan 12 22:35:08 <server> ntpdate[939]: step time server 198.72.72.10 offset -1.306097 sec
Jan 12 22:35:08 <server> ntpd:  succeeded
Jan 12 22:35:08 <server> ntpd[943]: ntpd 4.1.2 at 1.892 Tue Feb 24 06:32:25 EST 2004 (1)
Jan 12 22:35:08 <server> ntpd[943]: precision = 11 usec
Jan 12 22:35:08 <server> ntpd[943]: kernel time discipline status 0040
Jan 12 22:35:08 <server> ntpd: ntpd startup succeeded
Jan 12 22:38:24 <server> ntpd[943]: time set 0.000000 s
Jan 12 22:38:24 <server> ntpd[943]: synchronisation lost

I stopped ntpd again and have corrected the permissions on /etc/ntp and /etc/ntp/drift.  ntpd is now running again so I will see if this keeps it going.

/var/log/messages now has this:

Jan 12 22:49:11 <server> ntpd[1278]: ntpd 4.1.2 at 1.892 Tue Feb 24 06:32:25 EST 2004 (1)
Jan 12 22:49:11 <server> ntpd: ntpd startup succeeded
Jan 12 22:49:11 <server> ntpd[1278]: precision = 6 usec
Jan 12 22:49:11 <server> ntpd[1278]: kernel time discipline status 0040
Jan 12 22:49:11 <server> ntpd[1278]: frequency initialized 0.000 from /etc/ntp/drift

Hopefully the time will remain closer to correct now...

Again, any thoughts on this are welcome.

Thank you.
-Ryan

  ----- Original Message ----- 
  From: Adrin Story 
  To: Ryan Fish ; Atlanta Linux Enthusiasts 
  Sent: Wednesday, January 12, 2005 10:08 PM
  Subject: Re: [ale] Time issue


  Perhaps NASA has the answer already


NASA Science News for January 10, 2005 11:00:00 AM

The Dec. 26th Indonesian megathrust earthquake quickened Earth's rotation
and changed our planet's shape, according to calculations done by NASA
scientists.

FULL STORY at

http://science.nasa.gov/headlines/y2005/10jan_earthquake.htm?list64770

I had a machine at work that would pick up 8 seconds a day. (24 hours.)  You know this could be almost anything including soon to die hardware problem.


  Ryan Fish wrote:

    What could cause the time on a server running RHEL3 ES to change by roughly 3 hours at least two days in a row?

    - Time zone is set to America/New York (EST)
    - NTP is not in use
    - Box has not been rebooted for 77+ days (so HWClock should have nothing to do with it)

    Any ideas on where to look will be GREATLY appreciated.

    Thank you.
    -Ryan
----------------------------------------------------------------------------
_______________________________________________
Ale mailing list
Ale at ale.org
http://www.ale.org/mailman/listinfo/ale
-------------- next part --------------
An HTML attachment was scrubbed...