MiCollab Server (MAS) - NTP sync

[mpmcquade]

I recently started having problems with NTP sync on my MiCollab 7.1 server.  My other Mitel servers (2 - 3300's, vMCD, 2 - MBG's) all are syncing fine with an internal, Microsoft Windows 2012 R2 domain controller as the NTP time source.  My MiCollab server is pointing to the same IP address as the rest, but it shows .LOCL. as the 'refid' and does not have an asterisk (*) in front of the server name; it continues to gain time.  I changed it to manual mode and reset the time overnight as it was 3+ minutes ahead of the other servers, workstations and NTP domain controller source on the network.  I also rebooted the server after changing it back to automatic (use NTP) instead of manual.  This hasn't worked.  I can periodically query the NTP source from the Date and Time page in the management gui and see that it is gaining time.  Does anyone have any suggestions of how to correct this problem?  The only change in the environment was moving from windows server 2008 r2 domain controllers to windows 2012 r2 domain controllers which occurred several months ago.

[VinceWhirlwind]

The most obvious two questions are,
 - if the NTP server is identified by name, can the controller resolved that name?
 - can the controller reach the NTP server IP address on UDP port 123?

[mpmcquade]

Hello VinceWhirlwind,

the answer to your questions are yes to resolving the NTP server by name but I have tried using the IP address with the same results as well.  As far as the UDP port question, my assumption is yes to this also as there are no firewalls limiting access between the MAS server and the NTP server.  Querying the NTP server from the MAS gui returns continuous "YYYYY" in the reach column.

[mpmcquade]

I just ran a "nc -uzv <NTP servername> 123" command from the MAS console and received a "Connection to <ntp name> 123 port [udp/ntp] succeeded!" Attempted via IP address with the same results.

[acejavelin]

I have seen this happen before at one site... Changing the ntp server to a different, valid timeserver (like pool.ntp.org) and then back usually clears it up. Can't explain why though...

[Ericlaracuente]

Did you add the NTP server to the Micllab networks? I've seen that before but there was a bit of network segmentation between the servers, I added the NTP server address to the networks of the MiCollab server and ensured there were specific routes in our corporate firewall that permitted NTP to the Mitel server zones.

Sent from my Pixel 2 using Tapatalk