08-15-2018 07:32 AM
Hi ,
I have 4 NTP server configured on the switch, how does NTP server chosen among the 4. which value takes the precedence for server election. It keeps varying among four.
Which value is considered for election
Stratum
ref clock
delay
offset
08-15-2018 08:44 AM
You have to take many thing into account:
Before synchronizing, NTP compares the time reported by several network devices and does not synchronize with one that is significantly different, even if it is a stratum 1.
If all your servers have a similar reported time so the selection criteria will be based on the straum.
You can prefer a server: ntp server prefer (this in order to avoid your devices keep varying among the four NTPs)
Please do not forget to rate useful post.
Best Regards,
08-15-2018 11:09 PM
Dear Diana,
Thanks for your response, how to understand difference in reference time from below output. I haven't configured any prefer in my configuration. Before applying prefer command, need to decide which server is providing accurate time which is every closet to reference time
cisco#sh ntp associations
address ref clock st when poll reach delay offset disp
+~xx.250.113.149 xx.12.1.16 2 92 128 377 68.171 7.514 5.891
~xx.99.5.19 xx.99.20.200 3 115 128 1 1.753 6.745 187.55
*~xx.168.44.25 xx.71.80.83 1 22 128 377 0.971 7.150 3.891
~xx.99.5.10 10.9.xx.200 3 78 128 1 1.819 6.730 187.57
* sys.peer, # selected, + candidate, - outlyer, x falseticker, ~ configured
cisco#sh ntp status
Clock is synchronized, stratum 2, reference is xx.168.44.25
nominal freq is 250.0000 Hz, actual freq is 249.9972 Hz, precision is 2**22
ntp uptime is 622204100 (1/100 of seconds), resolution is 4016
reference time is DF1D601A.368FEB28 (00:05:14.213 AEST Wed Aug 15 2018)
clock offset is 7.1501 msec, root delay is 0.97 msec
root dispersion is 15.50 msec, peer dispersion is 3.89 msec
loopfilter state is 'CTRL' (Normal Controlled Loop), drift is 0.000011070 s/s
system poll interval is 128, last update was 429 sec ago.
08-15-2018 11:09 PM
Dear Diana,
Thanks for your response, how to understand difference in reference time from below output. I haven't configured any prefer in my configuration. Before applying prefer command, need to decide which server is providing accurate time which is every closet to reference time
Cisco# sh ntp associations
address ref clock st when poll reach delay offset disp
+~xx.250.113.149 xx.12.1.16 2 92 128 377 68.171 7.514 5.891
~xx.99.5.19 xx.99.20.200 3 115 128 1 1.753 6.745 187.55
*~xx.168.44.25 xx.71.80.83 1 22 128 377 0.971 7.150 3.891
~xx.99.5.10 10.9.xx.200 3 78 128 1 1.819 6.730 187.57
* sys.peer, # selected, + candidate, - outlyer, x falseticker, ~ configured
cisco#sh ntp status
Clock is synchronized, stratum 2, reference is xx.168.44.25
nominal freq is 250.0000 Hz, actual freq is 249.9972 Hz, precision is 2**22
ntp uptime is 622204100 (1/100 of seconds), resolution is 4016
reference time is DF1D601A.368FEB28 (00:05:14.213 AEST Wed Aug 15 2018)
clock offset is 7.1501 msec, root delay is 0.97 msec
root dispersion is 15.50 msec, peer dispersion is 3.89 msec
loopfilter state is 'CTRL' (Normal Controlled Loop), drift is 0.000011070 s/s
system poll interval is 128, last update was 429 sec ago.
08-15-2018 03:34 PM
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide