[lug] linux process supervision

D. Stimits stimits at idcomm.com
Thu Dec 7 19:39:49 MST 2000


Carl.Wagner at Level3.com wrote:
> 
> Doesn't nohup just prevent the process from dying when its parent process dies
> (like loging out or "hanging up").  I believe nohup stands for "no hang up".

Yes, but I would expect respawning should be in combination with nohup.
Running under nohup tends to make the app bail out only under more
serious conditions.

> 
> Carl.
> 
> "D. Stimits" wrote:
> >
> > "Ferdinand P. Schmid" wrote:
> > >
> > > HI,
> > > I wrote some scripts and a C++ application for one of my systems that I
> > > would like to keep running at all times.  During my tests for running these
> > > processes in the background I found that for some reason they sometimes go
> > > away (crash or die).  So here is my question:
> > >
> > > Services that are started out of inetd are watched out for by inetd (same
> > > thing for tcpserver).  So if they crash or die inetd restarts them
> > > immediately.  Is there a facility in Linux that I could use to take care of
> > > my own basic processes that are note internet related?  I would of course
> > > like to do this in a way that is not considered bad style - if there is
> > > such a thing.
> > >
> > > Thanks,
> > > Ferdinand
> > >
> >
> > Someone else already mentioned respawn. You will probably also be
> > interested in "nohup". Some shells have their own built-in of this, but
> > there is also a literal command nohup.
> >
> > _______________________________________________
> > Web Page:  http://lug.boulder.co.us
> > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> 
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug




More information about the LUG mailing list