On 11/30/05, Marco Peereboom <[EMAIL PROTECTED]> wrote: > Another very useful feature is the watchdog timer that is provided by the ESM > hardware. Watchdog timers need to be tickled every X seconds, if the > tickling doesn't happen it'll reboot the server. This is handy for servers > that are 1000 miles away and have unrecoverable kernel hangs (not rebootable > from ddb(4)). To enable the watchdog simply type: > # sysctl -w kern.watchdog.period=10 > > This will tickle the watchdog period/2, in this case every 5 seconds. Now if > the server hangs (e.g. put it in ddb(4)) it'll reboot after 10-ish seconds. > > To disable the watchdog simply set the period to 0, like so: > # sysctl -w kern.watchdog.period=0 > > As a sidenote I would like to mention that ipmi(4) has also been updated to > provide watchdog functionality.
Thank you, Jordan Hargrave and David Gwynne. Watchdog support is something I've been looking forward to for some time. This really is a huge win for deploying open on critical systems. Kevin
