Network Time Protocol (NTP)
You must install
NTP to
properly synchronize services among nodes. We recommend that you configure
the controller node to reference more accurate (lower stratum) servers and
other nodes to reference the controller node.
Controller node
To install the NTP service
# apt-get install ntp
# yum install ntp
# zypper install ntp
To configure the NTP service
By default, the controller node synchronizes the time via a pool
of public servers. However, you can optionally edit the
/etc/ntp.conf file to configure alternative
servers such as those provided by your organization.
Edit the /etc/ntp.conf file and add,
change, or remove the following keys as necessary for your
environment:
server NTP_SERVER iburst
restrict -4 default kod notrap nomodify
restrict -6 default kod notrap nomodify
Replace NTP_SERVER with the
hostname or IP address of a suitable more accurate (lower stratum)
NTP server. The configuration supports multiple
server keys.
For the restrict keys, you essentially
remove the nopeer and noquery
options.
Remove the /var/lib/ntp/ntp.conf.dhcp file
if it exists.
Restart the NTP service:
# service ntp restart
Start the NTP service and configure it to start when the system
boots:
# systemctl enable ntpd.service
# systemctl start ntpd.service
On SLES:
# service ntp start
# chkconfig ntp on
On openSUSE:
# systemctl enable ntp.service
# systemctl start ntp.service
Other nodes
To install the NTP service
# apt-get install ntp
# yum install ntp
# zypper install ntp
To configure the NTP service
Configure the network and compute nodes to reference the
controller node.
Edit the /etc/ntp.conf file:
Comment out or remove all but one server
key and change it to reference the controller node.
server controller iburst
Remove the /var/lib/ntp/ntp.conf.dhcp file
if it exists.
Restart the NTP service:
# service ntp restart
Start the NTP service and configure it to start when the system
boots:
# systemctl enable ntpd.service
# systemctl start ntpd.service
On SLES:
# service ntp start
# chkconfig ntp on
On openSUSE:
# systemctl enable ntp.service
# systemctl start ntp.service
Verify operation
We recommend that you verify NTP synchronization before proceeding
further. Some nodes, particularly those that reference the controller
node, can take several minutes to synchronize.
Run this command on the controller node:
# ntpq -c peers
remote refid st t when poll reach delay offset jitter
==============================================================================
*ntp-server1 192.0.2.11 2 u 169 1024 377 1.901 -0.611 5.483
+ntp-server2 192.0.2.12 2 u 887 1024 377 0.922 -0.246 2.864
Contents in the remote column should
indicate the hostname or IP address of one or more NTP servers.
Contents in the refid column typically
reference IP addresses of upstream servers.
Run this command on the controller node:
# ntpq -c assoc
ind assid status conf reach auth condition last_event cnt
===========================================================
1 20487 961a yes yes none sys.peer sys_peer 1
2 20488 941a yes yes none candidate sys_peer 1
Contents in the condition column should
indicate sys.peer for at least one server.
Run this command on all other nodes:
# ntpq -c peers
remote refid st t when poll reach delay offset jitter
==============================================================================
*controller 192.0.2.21 3 u 47 64 37 0.308 -0.251 0.079
Contents in the remote column should
indicate the hostname of the controller node.
Contents in the refid column typically
reference IP addresses of upstream servers.
Run this command on all other nodes:
# ntpq -c assoc
ind assid status conf reach auth condition last_event cnt
===========================================================
1 21181 963a yes yes none sys.peer sys_peer 3
Contents in the condition column should
indicate sys.peer.