home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!think.com!barmar
- From: barmar@think.com (Barry Margolin)
- Newsgroups: comp.unix.questions
- Subject: Re: <defunct> Zombies
- Date: 17 Dec 1992 20:24:38 GMT
- Organization: Thinking Machines Corporation, Cambridge MA, USA
- Lines: 26
- Message-ID: <1gqnm6INNnej@early-bird.think.com>
- References: <6JwuVB11w165w@iowegia.uucp> <29678@castle.ed.ac.uk>
- NNTP-Posting-Host: telecaster.think.com
-
- In article <29678@castle.ed.ac.uk> ajmy@festival.ed.ac.uk (A Myles) writes:
- >I've seen machines with week old zombies and no sign
- >of any living parent (i.e. a process matching PPID).
-
- If there are zombies whose parents are gone, then your system has a serious
- bug. Any time a parent process dies, its children are supposed to be
- inherited by init (pid 1). Init always calls wait() for any children it
- inherits, so a zombie should go away when its parent dies.
-
- > I don't think
- >its possible for a zombie to cling onto it's parent's parent is it,
- >while the "intermediate" process exits?
-
- It's not supposed to.
-
- >Is there some accounting purpose fulfilled by leaving zombies in
- >the table, or such like? (with and without parent)
-
- If there's a parent, the purpose of the process table entry is to hold the
- information that the parent can still access, such as the process's exit
- status (which is returned by the wait() call) and resource usage.
- --
- Barry Margolin
- System Manager, Thinking Machines Corp.
-
- barmar@think.com {uunet,harvard}!think!barmar
-