Skip to main content

Network Time Protocol (NTP): Windows versus Cisco

By April 3, 2018September 18th, 2020Blog, Cisco, Networking, Windows Server
cisco Network Time Protocol Diagram

I came across an issue where the IP phones were not able to keep proper time.  After troubleshooting the issue, I found that the DHCP option 42 (NTP server) pointed to the Windows 2016 server running windows time.  The phones couldn’t keep proper time.

I searched the Internet for issues with Windows and NTP and found blogs pointing to people having issues with NTP on windows.

  1. https://social.technet.microsoft.com/Forums/windowsserver/en-US/e0092a2b-54c6-4982-a6f4-a78ba53c0769/ntp-on-windows-server-2008-not-working-?forum=winserverDS
  2. https://social.technet.microsoft.com/Forums/en-US/22754f94-1af3-4c19-bc79-56a4d199ddfa/windows-ntp-server-does-not-appear-to-be-working?forum=winserverManagement

So it seems more difficult using time off the windows server than it is to setup NTP on the cisco switches.  So I opted to use NTP off the Cisco switches.

First, I swung the DHCP option to point to the Core switch by IP address.

Now typically I use NTP off the core network switch pointed back to the public pools at NIST: https://tf.nist.gov/tf-cgi/servers.cgi

Being in Boulder Colorado, I use the two University of Colorado NTP servers, but I haven’t had an issues with any of the NIST servers.

My configuration on my core switch is simple and easy once you have defined the servers you want to use.

CoreSwitch(config)#nntp server 128.138.140.44
CoreSwitch(config)#ntp server 128.138.141.172 prefer

To validate that NTP is working use the “show ntp status” command:

CoreSwitch #show ntp status
Clock is synchronized, stratum 2, reference is 128.138.141.172
nominal freq is 119.2092 Hz, actual freq is 119.2086 Hz, precision is 2**17
reference time is DE2C5A76.620E294F (11:24:54.383 MST Mon Feb 12 2018)
clock offset is -2.1640 msec, root delay is 5.14 msec
root dispersion is 28.62 msec, peer dispersion is 5.15 msec
loopfilter state is ‘CTRL’ (Normal Controlled Loop), drift is 0.000005685 s/s
system poll interval is 128, last update was 326 sec ago.

After validating NTP was functioning on the core switch and swinging over the DHCP option to point at the switch, we rebooted all the phones.  Didn’t have any issue with time again.  My point is that if you’re having issues with time, you probably want to just setup time on the network devices and point your clients to the switch.  It is much easier and stable than keeping time updated from the windows servers.

Jason Howe, PEI

Leave a Reply