Haven't done this myself but on researching the question because I was curious. I found that you basically need to add a logger line to the start and stop sections of each init script.
Examples I see in /etc/init.d/diskdump:
Code:
start_device() {
local dev=$1
errmsg=$($DISKDUMPCTL $dev 2>&1)
ret=$?
if [ $ret -ne 0 ]; then
logger -p info -t $SERVICE_NAME "$errmsg"
fi
}
stop_device() {
local dev=$1
errmsg=$($DISKDUMPCTL -u $dev 2>&1)
if [ $? -ne 0 ]; then
logger -p info -t $SERVICE_NAME "$errmsg"
fi
}
In the above basicaly it is saying to only log the information if there is a non-zero exit status (meaning it failed). You'd want to simply get rid of the if, then, fi so that it logged everything.
The OK/FAILED actually comes from the functions (see the file invoked at start of init scripts) so to actually see the OK/FAILED itself you'd likely have to modify that function script.
Unfortunately I didn't find any examples of anyone doing this.
Interestingly on HP-UX they have it setup to go to an rc.log by default but they do it differently than Linux.