The time for all nodes must be synchronized, which is usually achieved by configuring the NTP server. If the user already has a time server in the network, then all nodes can point to it, otherwise you can pick a node from the cluster as the time server and let the other nodes synchronize with it. These two methods are shown below:
(1) If a time server is already in the corporate network
Example: 192.168.11.10 is an NTP time server
At this point the NTP server's configuration file is/etc/ntp.conf, which is edited at each node.
[Email protected] ~]# vi/etc/ntp.conf
The edited content follows the preferred company's time server.
Server 192.168.11.10 prefer
Driftfile/var/lib/ntp/drift
Broadcastdelay 0.008
(2) If there is no external time server
When the company chooses a node in the cluster as a time server, the NTP service requires very little system resources. Assuming that the primary node is selected as a time server, synchronize it from the node and configure the method as follows.
Edit the main node's/etc/ntp.conf file, and edit the following content:
Server 127.127.1.0
Fudge 127.127.1.0 Stratum 11
Driftfile/var/lib/ntp/drift
Broadcastdelay 0.008
Edit the/etc/ntp.conf file from the node and edit the contents as follows:
Server 192.168.11.12 prefer #注意192.168.11.12 the IP address of the primary node
Driftfile/var/lib/ntp/drift
Broadcastdelay 0.008
After the configuration is complete, start the NTP service.
[[email protected] ~]#/etc/init.d/ntpd start
[Email protected] ~]# chkconfig–level 345 ntpd on
(3) View performance (master-slave)
[Email protected] ~]# ntpq-p
Remote refID St T when poll reach delay offset jitter
==============================================================================
*cc-node-02 LOCAL (0) u 58 64 17 0.226-22.698 29.261
Ext.: http://www.xifenfei.com/2011/07/oracle-10g-rac%e4%b9%8b%e9%85%8d%e7%bd%ae%e6%97%b6%e9%97%b4%e5%90%8c%e6%ad% A5%ef%bc%88ntp%ef%bc%89.html
Configuration time Synchronization (NTP) for Oracle 10g RAC