After installing Contrail 2.x, command contrail-status shows certain daemons in “initializing” state with the message “NTP state unsynchronized”. This article explains the issue and describes how to directly modify the /etc/ntp.conf file of all servers to configure a valid NTP server.

A snippet of contrail-status showing the problem :

During Contrail installation, it is necessary to configure a reachable NTP server in the testbed.py file. The installation scripts then configure this in / etc/ntp.conf after installation.

This is the section in testbed.py that should have the NTP server configured :

If this step is skipped during installation, it is still possible to fix this problem by directly modifying the /etc/ntp.conf of all servers to have a valid NTP server configured. Open the file /etc/ntp.conf in a vi editor and add the ntp server just below the following commented line :

Reboot the contrail nodes on which the above changes are made. It will take awhile for the servers to be NTP-synchronized.

Verification

Use the following command on the server to see if the server is NTP-synchronized. Contrail runs the ntpq command to check the NTP status.

ntpq -n -c pe | grep ^*

If you see an output with “*” at the beginning of the line, it means the Contrail server is synchronized successfully to a NTP server. For example,

If the command returns nothing, it means that NTP is unsynchronized. If you have configured the NTP server and still see the problem despite following the above guidelines, check the firewalls in between. Many of them, by default do not allow NTP traffic to pass by.