With our software application code installed on Linux Centos 5, we use the BMC watchdog ability with the pretimeout option to allow the software to get notified prior to the BMC rebooting the server. The BMC watchdog timer is set to expire in 300 seconds. We have seen on some servers for every 40 seconds we heartbeat with the BMC there is an entry in the BMC SEL
1fe | 01/05/2010 | 21:09:57 | Watchdog 2 #0x71 | Timer expired | Asserted
But the timer did not expire. Using "ipmitool mc watchdog get" shows the timer is set for 300 seconds and working just fine.
Any ideas why the SEL occurs every time we heart beat with the BMC watchdog ?
Again this is only on a few servers which are all configured the same.