Network topology :
the NTP server is integrated with windows 2012 NTP server, and installed in the virtual machine of E9000.
the issue description:
the NTP status of s9706 can synchonize, but after 2 minutes the status change to unsynchonized.
There are many warnings that system synchronization source lost.
Apr 6 2015 12:25:09+04:00 ADRAC1-CSW01 %%01NTP/4/SOURCE_LOST(l):System synchronization source lost. (SourceAddress=10.8.31.22, Reason=Clock selection failed - No selectable clock)
(1) ping successfully from the NTP client to the NTP server which excludes the situation that the network isn’t stable.
(2) Collect the debug information of ‘ntp-service all’, find the reason that Synchronization distance greater than distance threshold:
FileId [ 019 ], LineNo [ 00536 ]- Server 10.8.31.22 not considered for clock selection. (Reason: Synchronization distance greater than distance threshold. (Current: 11.00268, Allowable: 1.00048))
At the same time ,through the sync dist（11.075）of command ‘dis ntp sessions verbose’to further confirm the reason.
[switch]dis ntp sessions verbose
clock source: 10.8.31.22
reach: 377, sync dist: 11.075, sync state: 0
When use window software as the NTP server, the Synchronization distance deviates 10s from the default distance of S9706 (the same as others devices),because exceed the threshold, the NTP client will not synchronize the time of server, then fail to synchronize.
(1) Change the value of NTP server register ‘HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\W32Time\Config\LocalClockDispersion’ from 10 to 0
(2) Restart the NTP service by ‘w32tm /config /update’
Through that to guarantee that the deviation is within the Protocol scope.
When use windows software as the NTP server, the Synchronization distance deviates 10s from the default distance of huawei devices, because exceed the threshold, the NTP client will fail to synchronize. So had better not use windows software as the NTP server.