1.1.38 timeout before data read errors.

On Tue, 1 Jul 2003, Vernon Schryver wrote:
> > connect parameter?
> Correct parameter for what purpose?

connect! :)

> Did you read the comments in dcc.m4 file or the sendmail milter
> documentation in light of your needs?

no, but I did search the mailing list archive for this message and changed
this based on a earlier email you had replied too.


--- dcc-archive message below ---

Date: Wed, 22 Jan 2003 07:44:56 -0700 (MST)
From: Vernon Schryver <vjs@calcite.rhyolite.com>
Message-Id: <200301221444.h0MEiu0J012243@calcite.rhyolite.com>
To: DCC@rhyolite.com, lbolmerg@munichre.com
Subject: Re: Milter timeout message


> From: Bolmerg-Berliner Ludger - Munich-MR <lbolmerg@munichre.com>

> I have found a couple of entries like the following in my maillog file.
> Come somebody explain what they mean.  I am currently using dccm 1.1.16
and
> 1.1.17.  There is no local dccd running. dccm is using the public dccd
> servers.
>
> Jan 22 04:47:39 netserv2 sm-mta[52129]: h0M3lXGr052129: Milter (dcc):
> timeout before data read
> Jan 22 04:47:39 netserv2 sm-mta[52129]: h0M3lXGr052129: Milter (dcc): to
> error state

They mean that sendmail got tired of waiting for dccm to answer.

I think the default timeouts from misc/dcc.m4 in the DCC source are
5 and 30 seconds.  Perhaps the initial timeout should be increased
to 30 seconds.  That can be done by changing a line like
    Xdcc, S=unix:/var/run/dcc/dccm, F=T, T=C:5s;S:5s;R:5s;E:30s
to
    Xdcc, S=unix:/var/run/dcc/dccm, F=T, T=C:30s;S:5s;R:5s;E:30s
in sendmail.cf


Vernon Schryver    vjs@rhyolite.com


--
Mark Atkinson
darkmark@filament.org
(!wired)?(coffee++):(wired);