NTP not synchronizing in switches

[Sihanu N]

Hi Experts,

Recently we have configured NTP in our network environment. The current NTP setup is like following manner

Core1 switch(manually configured time) - ntp master

Core 2 swtch(manually configured time) - ntp master

In all Distribution and Access switches - ntp server <ip of core1 switch> prefer

                                                       ntp server <ip of core 2 switch>

Initially the synchronization is working fine. But recently its noted that the ntp is not synchronizing properly in some access switches after the reboot, but its detecting the ntp server.

The output of "show ntp associations" for unsynced access switches.

      address                                   ref clock          st  when poll reach  delay    offset      disp

~  <ip address of core1 switch>     127.127.7.1       8    55     64    0      1.2      -21870    16000.

~  <ip address of core2 switch>     127.127.7.1        8    59     64    0      1.2       21870.   16000.

* master (synced), # master (unsynced), + selected, - candidate, ~ configured

For some swithes its working fine.

1) What might be the unsyc problem in the access swithes, is it due to any mis-configuration in current setup?

Thanks and Regards,

Sihanu N

[Leo Laohoo]

But recently its noted that the ntp is not synchronizing properly in some access switches after the reboot, 

Are you saying that before the reboot the NTP sync was available?  If this is so, then it will depend on your IOS.  15.0(1)SE, SE1, SE2, takes exactly 10 minutes before NTP to sync.  Older IOS takes less time.  I don't know why this is so.

 but its detecting the ntp server. 

No it's not.  The IP Addresses of Core 1 and Core 2 have a "~" prefexing them.  This means that the NTP addresses are configured but are not or have not synchronized.

Can you please post your NTP command lines?

[Sihanu N]

Hi Leo,

Sorry for the delayed response and Many thanks for the reply to my post.

No, the issue shows as some access switches reloaded after the power outage. Once the same commands removed(using no command) are re-entered in the access swithces resolves the issue.

in access switches the commands given are as follows

ntp server prefer    

ntp server

Thanks & Regards

Sihanu N

[Leo Laohoo]

No, the issue shows as some access switches reloaded after the power outage.

Sorry, but can you explain this further?

In our network, our access switches (per site) gets NTP from the site distro.  The site distro gets their NTP from the core.  So when the site goes out due to power, the access swithces take 20 long minutes for the NTP to get sync.  This is because the distro takes 10 minutes for the NTP to sync.  During this stage the access switch NTP goes into insane.  10 minutes after the power outage the distro gets NTP sync but the access switch still doesn't get sync until a further 10 minutes.

I am not sure if this is what you are seeing.

What IOS versions are your switches running?