Well, for what it's worth, I was getting 'queue file write error' when I
didn't have enough mysql connections setup.  The errors seemed to be a
fairly even mix of 'queue file write error' and 'too many connections' (from
mysql).  After I increased the max connections so the queries could
complete, I didn't get any more 'queue file write error' messages either.

vec

----- Original Message -----
From: Wietse Venema <wietse <at> porcupine.org>
To: Postfix users <postfix-users <at> postfix.org>
Sent: Tuesday, May 14, 2002 10:57 AM
Subject: Re: 451 Error: queue file write error

> Really, how common is mail delivery that proceeds slower than 4096
> bytes per hour, while still sending something every 300 seconds?
>
> If this is a problem for you, specify
>
>     ipc_timeout = 7200s
>
> or something in main.cf.
>
> Wietse
>
> Charlie Watts:
> > On Mon, 13 May 2002, Frank Bonnet wrote:
> >
> > > I receive such error messages since I've installed 1.1.8
> > >
> > >  Out: 250 Ok
> > >  In:  DATA >  Out: 354 End data with <CR><LF>.<CR><LF>
> > >  Out: 451 Error: queue file write error
> > >  In:  QUIT
> > >  Out: 221 Bye
> > >
> > > what does this means ? postfix error ou FreeBSD error ?
> >
> > Vivek's suggestion is a good one - disk full. If you've verified that's
> > not it ...
> > ... I believe it's an esoteric timeout issue. Wietse made me a patch
that
> > logs an error when this occurs - below is part of the message from him,
I'm
> > fairly sure he won't mind my including it here.
> >
> > Wietse - the patch is working nicely for me. Every time I get one of
these 451
> > "error during smtp session" postmaster mails, there is a timeout logged.
> > Perhaps it would be appropriate to return some other error and not
> > generate a local postmaster notification?
> >
> > --
> > Charlie Watts
> > cewatts <at> frontier.net
> > Frontier Internet, Inc.
> > http://www.frontier.net/
> >
> > Wietse Venema wrote:
> > > The smtpd to cleanup socket has a read/write timeout of $ipc_timeout
> > > seconds, which is 3600 seconds by default.
> > >
> > > The remote client to smtpd socket has a read/write timeout of
> > > $smtpd_timeout seconds, which is 300 seconds by default.
> > >
> > > So, the client would have to send something every 300 seconds, but
> > > not enough to fill a 4096 byte I/O buffer for the smtpd to cleanup
> > > socket in 3600 seconds.
> > >
> > > This is possible in theory. I'll add a timeout warning to the
> > > cleanup server. See patch below.
> >
> > *** ./cleanup.c-        Mon Nov  5 11:25:58 2001
> > --- ./cleanup.c Wed May  8 16:19:54 2002
> > ***************
> > *** 232,237 ****
> > --- 232,244 ----
> >       }
> >
> >       /*
> > +      * Log something to make timeout errors easier to debug.
> > +      */
> > +     if (vstream_ftimeout(src))
> > +       msg_warn("%s: read timeout on %s",
> > +                state->queue_id, VSTREAM_PATH(src));
> > +
> > +     /*
> >        * Finish this message, and report the result status to the
client.
> >        */
> >       attr_print(src, ATTR_FLAG_NONE,
> >
> >
> >
> > -
> > To unsubscribe, send mail to majordomo <at> postfix.org with content
> > (not subject): unsubscribe postfix-users
> >
> >
>
> -
> To unsubscribe, send mail to majordomo <at> postfix.org with content
> (not subject): unsubscribe postfix-users
>

-
To unsubscribe, send mail to majordomo <at> postfix.org with content
(not subject): unsubscribe postfix-users