Time differential between IP phones and computer

[norman.west]

We're running CUCM 6.0. Users are reporting a time differential between their IP phone and their computer/cell phone. The publisher is referenced to a NTP source and is OK. The problem is that the subscriber is not synchronized with the publisher, confirmed by ssh'ing into each server and using the command "utils ntp status". Is there a way to synchronize them or is this a bug?

[htluo]

Could you SSH into the publisher and type "utils ntp status"? Then copy/paste the screen output here.

After that, SSH into the subscriber and type "utils ntp status". Then copy/paste the screen output here.

Thanks!

Michael

http://htluo.blogspot.com

[norman.west]

Thank you Mike.

PUB:

admin:utils ntp status

ntpd (pid 6198) is running...

remote refid st t when poll reach delay offset jitter

==============================================================================

127.127.1.0 127.127.1.0 10 l 52 64 377 0.000 0.000 0.004

+149.179.1.29 172.24.246.14 5 u 2 64 377 53.125 3307.36 32.633

*149.179.1.30 127.127.7.1 2 u 3 64 377 61.318 -1216.4 32.312

synchronised to NTP server (149.179.1.30) at stratum 3

time correct to within 4568 ms

polling server every 64 s

Current time in UTC is : Wed Sep 2 01:53:16 UTC 2009

Current time in America/Denver is : Tue Sep 1 19:53:16 MDT 2009

=======================================

SUB:

admin:utils ntp status

ntpd (pid 5884) is running...

remote refid st t when poll reach delay offset jitter

==============================================================================

*127.127.1.0 127.127.1.0 10 l 50 64 377 0.000 0.000 0.004

138.241.136.10 149.179.1.30 3 u 53 64 377 0.169 1409793 2.945

synchronised to local net at stratum 11

time correct to within 11 ms

polling server every 64 s

Current time in UTC is : Wed Sep 2 01:29:56 UTC 2009

Current time in America/Denver is : Tue Sep 1 19:29:56 MDT 2009

[htluo]

How long the subscriber NTP has been in this status?

Due to the design of NTP protocol, the subscriber won't synchronize its time with publisher immediately.

If it took more than an hour to sync up, you should get NTP logs.

Michael

[norman.west]

1) The problem was reported in mid-August. It's cropped up before in the March/April time frame. (I think) the cluster was rebooted at that time and that seems to have resolved it. However it's probably been drifting since then.

(2) What is the CLI syntax to obtain the ntp.log file? I'm logged into the PUB and execute the command:

"admin:file list activelog syslog" yields

CiscoSyslog CiscoSyslog.1

CiscoSyslog.2 CiscoSyslog.3

CiscoSyslog.4 boot.log

boot.log.ori cron

cron.1 csalog

csalog.0 csalog.1

csalog.2 csalog.3

maillog messages

messages.ori ntp_one_shot.sh.log

ntp_one_shot.sh.log.0 ntp_reachability.sh.log

ntp_start.sh.log ntp_stop.sh.log

ntpd.log sd_ntp.log

sd_ntp.log.0 secure

secure.1 secure.2

secure.3 secure.4

secure.ori securitylog.txt

spooler

dir count = 0, file count = 33

How do I get to view the content of the ntp file?

"admin:file view activelog syslog /sd_ntp.log" yields

Expected 1 mandatory and up to 0 non-mandatory parmeter(s)

but 2 parameter(s) were found

[j.huizinga]

maybe you already figured out, but you have to put a / before syslog

JH

[j.huizinga]

stratum 11 is the problem.

You must stay below 10

JH