From owner-freebsd-questions Thu Jun 27 18:05:18 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id SAA13105 for questions-outgoing; Thu, 27 Jun 1996 18:05:18 -0700 (PDT) Received: from salsa.habaneros.com ([207.34.140.99]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id SAA13100 for ; Thu, 27 Jun 1996 18:05:15 -0700 (PDT) Received: from jalapeno.habaneros.com (jalapeno [207.34.140.98]) by salsa.habaneros.com (8.6.12/8.6.12) with SMTP id RAA07813; Thu, 27 Jun 1996 17:49:45 -0700 Received: by jalapeno.habaneros.com with Microsoft Mail id <01BB6452.4DBABC40@jalapeno.habaneros.com>; Thu, 27 Jun 1996 17:58:53 -0700 Message-ID: <01BB6452.4DBABC40@jalapeno.habaneros.com> From: "Neil C. Jensen" To: James Raynard , "'Information Help Desk'" Cc: "freebsd-questions@FreeBSD.ORG" Subject: RE: trouble for process to be terminated Date: Thu, 27 Jun 1996 17:58:49 -0700 Encoding: 80 TEXT Sender: owner-questions@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk I have a similiar problem with serial io on COM2, although I am using getty: ttyd1 "/usr/libexec/getty std.19200" dialup on In my case, I can dialup, and login once ok. However, when I exit from getty, the process and user shell refuses to die. My next attempt to dialup always fails, the modem picks up and then disconnects. This kills the original process and starts a new one, and I can then log in again. This is repeatable, successful login, unsuccessful, successful..... I haven't tried in on COM1, but now I think I will... ---------- From: Information Help Desk[SMTP:info@adn.edu.ph] Sent: Friday, June 28, 1996 1:40 AM To: James Raynard Cc: freebsd-questions@FreeBSD.ORG Subject: Re: trouble for process to be terminated On Thu, 27 Jun 1996, James Raynard wrote: > > What's the last resort besides rebooting after killing a process > > that still lives but is trying to exit? > > Are you referring to a "zombie" process - one where a child process > exits before its parent and the parent doesn't clean up after it > properly (by calling wait())? > > It's not as bad as it seems, as the process has actually exited, > releasing all the memory it was using and closing any files it had > open, etc. All it's taking up is an entry in the process table. On the > other hand, it effectively reduces the number of processes you can run > by one, and makes the ps output look untidy. > > This is usually the result of a programming mistake - or the parent > may be blocking on a resource which never becomes available. Perhaps > if you can post some details about which program causes the problem, > and how it happens, we may be able to give a more helpful answer. > > -- > James Raynard, Edinburgh, Scotland > james@jraynard.demon.co.uk > HI !!! The process here is mgetty which I used for dial-up and whose owner is the init process. In my /etc/ttys, I have an entry, ttyd0 "/usr/local/sbin/mgetty -Ds 19200 -m'""AT&H1 OK'" dialup on secure This works just fine and I have the modem connected to COM1. You might find it weird that instead of cuaa0, I used ttyd0. Somebody told me cuaa0 should be used but I was able to make it work with using ttyd0. This setup works just fine. After every connection and disconnection, the process mgetty is terminated and resurrected by the init process. Now, I encounter this problem with the process mgetty not being able to exit when instead of ttyd0, I used ttyd1 and have the modem connected to COM2. I made the necessary changes in /etc/ttys, turned the line off in the dial-up in COM1, and added a modified line similar to the above. After the changes, I now see a process started by init. I tried to issue a 'kill -9 '. What should happen is the process exits and gets resurrected. But, the process seems having problems exiting and I usually see the line below. ???? ?? IE ?:??.?? /usr/local/sbin/mgetty -Ds 19200 -m AT&H1 OK ttyd0 I looked this up and this means that the process knows it has been issued a SIGKILL but still is trying to exit. If details are still needed I am more than willing to replay ASAP. Thank you. -- jf