[BusyBox] httpd and init

[Paul Fox]

 > 
 > > Does 
 > > no one ever check to see if httpd or any daemon has
 > > exited when you don't want it to?
 > 
 > Not really, no.  If that happens, you generally notice because
 > the service in question has gone down.  (In the case of a web
 > server, it's eminently noticeable.) And WHEN that happens, you
 > want to know WHY.

we've been through this on this list before, both for httpd and
dhcpd, both in the last year.  not everyone agrees on these
points, and answers aren't the same for all types of systems. 

sure it's great to know why a daemon is dying, so someone can do
something about, but only if there _is_ a someone to do something
about it.  if restarting it will keep my $200 appliance keep
running after it got an odd malformed packet that the coder
didn't account for, then restarting the daemon is probably what
the customer wants.  put another way:  as a developer, i can look
at a log and see that a critical service is respawning once in a
while, or i can look at a log and see...  nothing.  if the box
manages to ship that way, in one case the customer has service,
and in the second case...  nothing.

summary:  there are patches for an httpd foreground option in the
list archives, if you're interested.

paul
=---------------------
 paul fox, pgf at brightstareng.com