[ntp:questions] sync state and time to get there.

Larry Brigman larry.brigman at gmail.com
Tue Jun 23 05:57:31 UTC 2009

I am working with quite a few servers that have a time critical streaming
app running on a Linux ( x86_64).
These apps have no tolerance for a jump in time like a time reset so we have
set the -g -x options to ntpd.
We have a window of time before the streaming apps start to set the time and
get into time sync.

The problem I am seeing on start up of these machines (w/o drift file) that
it takes 24-72 hours to get to a sync state
but when the server is loaded prior to the sync state it never seem to get
to the sync state.  We have some machine
that have been running 12 days still without getting to ntp sync state.

With the below info, what could I do to improve the process of getting to a
sync state faster but not get time

Config file
restrict default kod nomodify notrap nopeer noquery
restrict -6 default kod nomodify notrap nopeer noquery

server ntp_server burst iburst

server     # local clock
fudge stratum 10

driftfile /var/lib/ntp/drift
keys /etc/ntp/keys

[root at localhost ~]# ntpdc -c kerninfo
pll offset:           0 s
pll frequency:        0.000 ppm
maximum error:        2.00952 s
estimated error:      1.6e-05 s
status:               0040  unsync
pll time constant:    4
precision:            1e-06 s
frequency tolerance:  500 ppm
[root at localhost ~]# ntpdc -c loopinfo
offset:               0.000197 s
frequency:            -68.482 ppm
poll adjust:          28
watchdog timer:       635 s
[root at localhost ~]# ntpq -pn
     remote           refid      st t when poll reach   delay   offset
*     4 u   80  128  377    0.288    0.889

Startup options
 7753 ?        SLs    0:00 ntpd -g -x -u ntp:ntp -p /var/run/ntpd.pid

More information about the questions mailing list