Difference between revisions of "Machine reboot without very good reason"
m |
m |
||
| Line 1: | Line 1: | ||
<yambe:breadcrumb> | <yambe:breadcrumb>Server administration tips and tricks</yambe:breadcrumb> | ||
=Machine reboot without very good reason= | =Machine reboot without very good reason= | ||
Revision as of 13:09, 15 January 2017
<yambe:breadcrumb>Server administration tips and tricks</yambe:breadcrumb>
Machine reboot without very good reason
CentOS machines come with service bmc-watchdog which sets up a watchdog timer. Whenever this timer reaches zero it assumes machine has crashed and resets the machine. The service which resets the timer value before it become zero is 'ipmi'. Thus if ipmi service is disabled and bmc-watchdog is not then machine will reboot frequently without any reason. To solve the problem enable ipmi service rather than disabling bmc-watchdog. If machine is configured to start appropriate services, VMs and containers on reboot then it is good to have bmc-watchdog to reset machine, if it has really crashed.
Learned from http://www.quantaqct.com/en/02_support/03_faq_detail.php?mid=71
