CUCM NTP add/delete/modify

[will.alvord]

On the NTP server list page under OS admin, the help for this page lists this:

"After deleting, modifying, or adding the NTP server, you must restart  all other nodes in the cluster for the changes to take affect."

I had never seen this before, but am troubleshooting an NTP issue and want to verify that a cluster reboot isn't needed.

thanks,

will

[Rob Huffman]

Hi Will,

It's true

NTP Servers

Ensure  that external NTP server is stratum 9 or higher (1-9). To add, delete,  or modify an external NTP server, follow this procedure:

Note You can only configure the NTP server settings on the first node or publisher.

Procedure

Step 1 From the Cisco Unified Communications Operating System Administration window, navigate to Settings>NTP Servers.

The NTP Server Settings window displays.

Step 2 You can add, delete, or modify an NTP server:

•To delete an NTP server, check the check box in front of the appropriate server and click Delete.

•To add an NTP server, click Add, enter the hostname or IP address, and then click Save.

•To modify an NTP server, click the IP address, modify the hostname or IP address, and then click Save.

Note  Any change that you make to the NTP servers can take up to 5 minutes to  complete. Whenever you make any change to the NTP servers, you must  refresh the window to display the correct status.

Step 3 To refresh the NTP Server Settings window and display the correct status, choose Settings>NTP.

Note  After deleting, modifying, or adding the NTP server, you must restart  all other nodes in the cluster for the changes to take affect.

http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/cucos/7_0_1/cucos/iptpch4.html#wp1042131

Cheers!

Rob

"Far away from your trouble and worry
You belong somewhere you feel free" - Tom Petty

[will.alvord]

It's true that it's listed in the docs, but not sure it's actually true.    That's what I wanted to confirm.  Since the subs sync from the publisher, it doesn't make sense that rebooting the subs (since it implies only the subs need a reboot) would be needed.  If anything, I can see the pub needing a reboot since the only change is being made to it.

I ran a packet capture on the pub filtering only on port 123, made the change, and stopped the capture.  I can clearly see that the change took affect immediately.  Since at this point nothing changes on the subs (they're still pulling ntp from the pub), it makes no sense at all that the subs would need a reboot right?

[Rob Huffman]

Hi Will,

I linked the doc as it reflects the "best practices" for CUCM

If it were me...I would reboot...but I did find this thread as well

https://supportforums.cisco.com/thread/2059615

Cheers!

Rob

"Far away from your trouble and worry
You belong somewhere you feel free" - Tom Petty

[will.alvord]

Just wasn't sure if that was a typo meant to advise a 'utils ntp restart' rather than a server restart.  Curious though if you typically reboot the pub as well or just the subs?

[Rob Huffman]

Hey Will,

I'm a pretty paranoid person

I usually take the opportunity to reboot the whole cluster. Having

the documentation makes it easier for us to get apporovals

from the Change Management peeps so I prefer the "safe vs sorry"

approach....

Cheers!

Rob

"Far away from your trouble and worry
You belong somewhere you feel free" - Tom Petty

[will.alvord]

I hear you.  Thanks Rob!

[Rob Huffman]

Hi Will,

You are most welcome my friend

Cheers!

Rob

"Far away from your trouble and worry
You belong somewhere you feel free" - Tom Petty