Re: [VANILLA-LIST:3115] Monitoring Netrek Services

[Dave Ahn <ahn@vec.wfubmc.edu>]

On Thu, May 04, 2000 at 09:53:55AM +1000, James Cameron wrote:
> 
> Actually, if netrekd started daemonII instead of ntserv, then it could
> wait on the SIGCHLD it gets in order to know if it is alive or not.
> 
> So far we have resisted any shared memory access within netrekd, for
> fairly good reasons.  Currently netrekd is fully restartable; it is
> possible to break this feature with lack of care.  ;-}

Actually, you wouldn't need shm access for newstartd.  The easiest way
I can think of is to let daemonII send SIGALRM to newstartd regularly
where it can be handled to keep track of a "daemonbust" counter.  Given
sufficiently high value, newstartd could kill daemonII (if it exists)
and restart it.

SIGCHLD would be cleaner, though.

-- 
Dave Ahn <ahn@vec.wfubmc.edu>        |  "When you were born, you cried and the
                                     |  world rejoiced.  Try to live your life
Virtual Endoscopy Center             |  so that when you die, you will rejoice
Wake Forest Univ. School of Medicine |  and the world will cry."  -1/2 jj^2