Processing large email

[Tim Sailer]

I appreciate all the good suggestions, both on and offlist, but,
like I said in the email, we're running Exim, and switching to
sendmail is not a option.

Tim

On Tue, Feb 04, 2003 at 01:28:58PM -0500, Jeff A. Earickson wrote:
> Hi,
>
>    I suggest setting a maximum message size in sendmail.  The
> m4 entry is:
>
> define(`confMAX_MESSAGE_SIZE', `10485760')dnl
>
> for a 10MB maximum, for instance.
>
> The third edition of the Bat Book (O'Reilly's sendmail book) has
> a good write-up about configuring multiple sendmail queues and
> multiple .cf files to handle queues of different size/speeds.
> They give an example of a "fast" queue and a "slow" queue.
> Every email gets one chance to get processed with the fast queue
> options and if it doesn't go out the first time, then it ends up
> in the "slow" queue for later retries.  I read thru the chapter
> when I had the Bat Book in hand at a book store; I've ordered a
> copy but don't have it yet.
>
> --- Jeff Earickson
>
> On Tue, 4 Feb 2003, Tim Sailer wrote:
>
> > Date: Tue, 4 Feb 2003 12:22:13 -0500
> > From: Tim Sailer <sailer at BNL.GOV>
> > Reply-To: MailScanner mailing list <MAILSCANNER at JISCMAIL.AC.UK>
> > To: MAILSCANNER at JISCMAIL.AC.UK
> > Subject: Re: Processing large email
> >
> > On Tue, Feb 04, 2003 at 05:15:34PM +0000, Julian Field wrote:
> > > At 17:02 04/02/2003, you wrote:
> > > >I've been using mailscanner with great success for a while, but
> > > >every once in a while, the mail queue seems to jam up. Investigation
> > > >shows that it's large emails, either attachments, or just text,
> > > >but 100MB+. Is there any way (running exim) to either push this
> > > >off to another queue, or give the large email a low priority, so
> > > >it doesn't get scanned until there's nothing left to do?
> > >
> > > A large email will only block up one of the child processes. The others
> > > will continue scanning other mail. If you run with the default 5 children,
> > > then you would have to have 5 huge mails turn up at *exactly* the right
> > > moments so as to block all 5 children.
> >
> > It happens, trust me. I'm graphing mail latency, and sometimes the
> > latency goes up to 40+ minutes. :(
> >
> > Tim
> >
> > --
> > Tim Sailer <sailer at bnl.gov>
> > Application Services
> > Information Technology Division
> > Brookhaven National Laboratory  (631) 344-3001
> >
>

--
Tim Sailer <sailer at bnl.gov>
Application Services
Information Technology Division
Brookhaven National Laboratory  (631) 344-3001