[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Immortal processes



On Monday 09 January 2006 01:32 pm, Jan C. Nordholz wrote:
> Hi Marc,
>
> > I am having a similar problem.  I keep my box on 24x7 and X is up the
> > whole time.  I have a window devoted to TkRemind to provide me with a
> > calendar.  It also pops up a reminder window each night at midnight with
> > the next day's reminders.  The problem is that it runs remind each time
> > to get the new reminders and then it leaves the zombie process behind,
> > and it can not be killed, even with a signal 9.  (Even as root, I can
> > not get rid of these processes.)
>
> a zombie process is a different thing. Zombie processes don't exist
> anymore, and their memory has already been freed. The only thing left is
> the return code, which the parent process can retrieve via C's wait(),
> waitpid() and friends. As long as the father process doesn't recollect its
> dead children (or terminates itself), those zombies will remain and,
> because they're no real processes anymore, not respond to signals.
>
> So summing up, this sounds like a bug in TkRemind... where exactly is
> TkRemind? Could you name me the package? (I'd go look myself, but
> packages.d.o is currently down...)

I don't know what language TKRemind is written in, but if it is in Perl, the 
fix should be easy enough.  It just needs a variable set to ignore zombies.

Hal



Reply to: