Recently, we had a hardware failure on our NTP server which caused its crash. Due to this, the ntpd daemon on compute nodes and cell nodes went offline. After fixing the hardware issue on NTP server, starting ntpd service using "service ntpd start" on compute nodes was successful but when I started the ntp service on cell node, it rebooted !!!. Luckily, cell node came up fine after the reboot and was connected to ntp server, But the question is, why it got rebooted!! and what is the correct way to restart ntpd service or change NTP server.
I am thinking that, to maintain cluster integrity it went for a reboot, but not sure exactly about it. Oracle has a comprehensive documentation on how to modify ntp safely on entire Exadata machine so as to avoid cell node reboot. Following are the references of the correct procedure:
How to make Network Time Protocol (NTP) changes on Compute and Cell nodes without node reboot (Doc ID 1963463.1)
Steps to shut down or reboot an Exadata storage cell without affecting ASM (Doc ID 1188080.1)
Above references should be followed before making any NTP changes to Exadata machine.