Well, this is how I understood the documentation:
When running from inetd, the mail spool file is copied and the popper spools
from that file. When running in standalone mode that doesn't happen.
Therefore, the reasoning behind choosing standalone is that the machine is a
bag of rubbish really for the task. (AMD K6 300MHz with only 1 physical IDE
disk although 512 MB of RAM)
Since IDE requires not insignificant CPU overhead, I don't really want to
copy a large mailbox (many customers have mailboxes > 30-40 MB) each time
they log in.
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: 13 December 2000 14:08
To: Peter R. Hubberstey
Subject: Re: defunct processes in server mode ???
Hi Peter
What was the problem with running it from inetd ?
We do that here at Moonlight without problem (and also with the benefit of
tcp wrappers).
Peter
At 12:08 13/12/2000 +0000, you wrote:
>Hello, my name is Peter Hubberstey and I work for a small ISP in the UK.
>
>I have compiled a new version of QPOP (version 3.1) some time ago and run
it
>bound to port 2000. I informed a few of our technically literate users and
>asked if they could test it.
>
>The testing has gone OK for some months now ( about 3 ) so last night I
>decided to change the main version.
>
>I decided to rem the line from the /etc/inetd.conf and restart the inetd
>daemon and run popper in stand alone mode instead. We have about two
>thousand customers using this server and whilst probably a maximum of
around
>10 or so ever collect mail at any one time, the disks are IDE so I figured
>that popper's server mode would be a better choice.
>
>So I run the daemon.
>
>/usr/local/popper -SscT 120
>
>All Fine.
>
>Checked this morning and I know have loads of "defunct" task
>check out the ps -ef
>
>root 28598 1 0 Dec12 ? 00:00:02 /usr/sbin/popper 110 -SscT
>120
>thr 24270 28598 0 08:59 ? 00:00:00 [popper <defunct>]
>smf 29619 28598 0 10:23 ? 00:00:02 [popper <defunct>]
>rapgroup 30558 28598 1 10:38 ? 00:00:34 [popper <defunct>]
>srbc 31274 28598 1 10:49 ? 00:00:27 /usr/sbin/popper 110 -SscT
>120
>jreeves 31847 28598 0 10:58 ? 00:00:00 [popper <defunct>]
>deltek 821 28598 0 11:18 ? 00:00:00 [popper <defunct>]
>pcmg 946 28598 0 11:21 ? 00:00:00 /usr/sbin/popper 110 -SscT
>120
>root 1004 28598 0 11:21 ? 00:00:00 /usr/sbin/popper 110 -SscT
>120
>rapgroup 1094 28598 87 11:23 ? 00:00:12 /usr/sbin/popper 110 -SscT
>120
>
>Trouble is I have found that by killing the defunct tasks, it brings down
>the main daemon too.
>
>Any known problems? Suggestions ?
>
>RedHat 6.2 on an AMD K6 300 using glibc 2.1.1-6
>
>Regards;
>
>Peter (UKIP)
>
