I am also getting this behaviour, and I'm not sure why.  I'm not using
supervise, instead I'm simply using --detach.  I thought it might be because
the occasional bombout with smtp-forward. (Getting Net::Cmd errors) Some of
the forks are simply not
 terminating.  I'm not using clamav or spamassasin (yet), but I do have most
of the 'default' plugins running.

Jan



On 12/28/05, Bob Dodds <[EMAIL PROTECTED]> wrote:
>
> Juerd wrote:
> > Skaag Argonius kribis 2005-12-28 15:23 (+0200):
> >
> >> But I have a feeling something is wrong. All the other 35 connections
> >> are stuck, they do not complete at any time. That line keeps coming
> back
> >> every second forever.
> >>
> >
> > The message repeats until there are free connection slots again.
> >
> >
> >> What do you say?
> >>
> >
> > I think you should use strace or something like it to find out what the
> > processes are waiting. Don't exclude the possibility of a malicious
> > client that just keeps connections open.
> >
> >
> > Juerd
> >
> Show us your ./run file. Weird things can happen in there.
> I just found out yesterday that clamd was restarting every
> second because the freshclam update in clamav/run was
> set up wrong.
>
> What is your last log line for each connection(not these?)--
>
> @4000000043b19a4b1b10fa5c 24652 Initializing spool_dir
>
> 000000043b19a3e1cbbef24 Error creating server 0.0.0.0:2525 :
> IO::Socket::INET: Address already in use
>
> Socket in use could mean that there's a zombied instance
> of tcpserver/qpsmtpd on port 2525. kill -9 stuff or reboot.
> You'd have to kill -9 the supervise process first, if it won't
> go away by ha ha ha "svc -d /service/qpsmtpd".
>
> -Bob
>
>
>

Reply via email to