SUMMARY: who's locking that file (revisited)?

From: Frank Cusack (fcusack@voicenet.com)
Date: Mon Dec 08 1997 - 11:57:28 CST


Thanks to Casper Dik <casper@holland.Sun.COM> (the only respondent) for
his reply (why would init not reap a zombie process):

>
> If the process is closing devices, it may remain exiting.
> Lsof is typically able to show the show the files still open.
>
>
> There are also bugs in older releases of Solaris that may cause init
> not to reap processes
>
>

I want to add that the times I have seen a zombie with PPID of 1 have
all been with processes doing serial port comm.

The original question:

>
> >My first try to post this didn't seem to work:
> >
> >In a previous summary:
> >
> >> > Question:
> >> > Why would a defunct prcess have a ppid of 1 ?
> >> > Answer:
> >> > Once a process becomes a zombie, its parent ID reverts to that of
> >> > init, which is 1.
> >> >
> >>
> >> Not quite right. Once a process becomes a zombie, it hangs around until
> >> the parent reaps it. The PPID should still be that of the original parent.
> >> If the parent exits, THEN the PPID becomes 1. At which point init
> >> should reap it.
> >>
> >> So the question is unanswered: why would init not reap a defunct process?
> >> Or maybe, why would wait() not return if the process has exit'ed?
> >>
> >> I've seen this once, there was a defunct process with PPID 1, and
> >> another with PPID of the first defunct process. It shouldn't matter
> >> that the zombie also has children, should it?
> >
> >
>
>

-- 
~frank
* I am Pentium of Borg. Division is futile. You will be approximated. *
*        PGP ID: C001AA75         -|-      fcusack@voicenet.com       *



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:11 CDT