4.26- beta upgrade (was RE: Another MailScanner User!)

[Julian Field]

Any chance you could please take a copy of the current Postfix docs off the
MailScanner site, update them to use the "hold queue" approach, and mail
the resulting docs back to me so I can put them up on the web site (at
least as an alternative, quite possibly as a replacement).

This all sounds very good to me. I partially did it the way I did so that
the setup was the same split-MTA as sendmail, but if this route is more
reliable then I think people should be using it.

At 11:51 20/01/2004, you wrote:
>As I understand Postfix doesn't use much in the way of file locking. It
>doesn't need to. In standard form a message is dropped into the onward
>directory and the next process is called using a 1b message and so mail
>makes it's way through the MTA. MailScanner upsets it by trying to grab
>the file from the deferred directory for processing. Now the deferred
>directory is used by Postfix as the place where mail is put when delivery
>fails, pending re-try (Keeps the active queues down) and every so often
>(As set in master.cf) the queue runner process goes to the deffered queue
>and inspects the messages for any that are due for retry. If the time
>stamp has expired it picks up the message and trys to deliver it. Through
>all of this there is not a need for much in the way of locking as what is
>going to touch that file? Postfix (As far as Postfix is concerned!) and
>Postfix knows what it's doing (We hope :-) ) If MailScanner and Postfix
>queue runner should happen to try to take the same message, you get the
>'still being delivered' message in the logs and up pops a duplicated mail!
>
>Easy way round it, use the hold queue. This is designed to only have
>messages dropped in it for leter inspection by the postmaster and so the
>queue runner doesn't ever re-inspect this directory. Ideal for
>MailScanner, message gets dropped (MS knows how to tell when it's
>complete), picks up the new message, does it's bit and puts it back in the
>incoming queue for Postfix to deal with in it's usual efficient manner.
>
>I haven't had a single duplicate since putting this in place.
>
>Drew
>
>Neil Robst said:
> > And you think this resolves the duplicate mail problem?
> >
> > I'm unsure how it differs (apart from only having one postfix daemon
> > running) from using /var/spool/postfix.in/deferred and
> > /var/spool/postfix/incoming...?
> >
> > However, as I've just had a report from my users saying that upgrading
> > to the 4.26-4 beta hasn't worked, I'm willing to try anything :-)
> >
> >
> > On Tue, 2004-01-20 at 10:20, Drew Marshall wrote:
> >> Neil
> >>
> >> What I have done is below, as suggested by Peter Bates and forwarded to
> >> me
> >> from this list.
> >>
> >> > I'm using MS with Postfix in a slightly 'non-standard' way, but which
> >> is
> >> working fine for 13-15K messages we deal with (actually it might be
> >> more, I never bothered counting our outgoing email!)...
> >> > I'm using a 'header_check' like so:
> >> > In main.cf -
> >> > header_checks = pcre:/etc/postfix/header_checks
> >> > In header_checks -
> >> > /^Received:.*by .*\.your.domain.tld \(Postfix\)/ HOLD
> >> > This puts the incoming mail in the 'hold' queue, and then
> >> > I have in MailScanner.conf -
> >> > Incoming Queue Dir = /var/spool/postfix/hold
> >> > Outgoing Queue Dir = /var/spool/postfix/incoming
> >>
> >> With this, you will need to stop postfix.in and uncomment the smtp line
> >> in
> >> master.cf (Basically revert your set up to a non-MailScanner set up (It
> >> may be easier if Postfix.in runs chrooted and postfix doesn't to just
> >> alter postfix.in to become just postfix, what ever your mileage!)). Stop
> >> all instances and restart just postfix and you now have one postfix
> >> instance with MailScanner.
> >>
> >> Works great!
> >>
> >> Drew
> >> --
> >>
> >>
> >> Neil Robst said:
> >> > Drew,
> >> >
> >> > Can you explain a bit more about how you've configured postfix,
> >> please?
> >> I'm using the suggested setup of two postfix instances - the first runs
> >> everything in a chroot jail and smtp, local and virtual and deferred.
> >> Mailscanner then picks everything out the deferred queue, does it's
> >> stuff and drops it back into the incoming queue of the second postfix
> >> instance. Seems to be working well, but you said you'd changed postfix
> >> to bypass the duplicate problems...
> >> >
> >> > Regards,
> >> > Neil
> >> >
> >> > On Tue, 2004-01-20 at 09:06, Drew Marshall wrote:
> >> >> I've been running it now since the weekend without problem. I would
> >> suggest that although marked as a beta and potentialy unstable, it's
> >> about
> >> >> as unstable as the production releases :-) The new patches seem to be
> >> working well.
> >> >> I have to admit, I changed my Postfix set up to by pass the duplicate
> >> problems and haven't changed it back. I now use a rule in Postfix to
> >> hold
> >> >> all incoming mail, let MS collect from the hold queue (The queue
> >> runner
> >> doesn't ever run in there) and drop back into the incoming queue for
> >> delivery. It just means that I only have to ever run just one Postfix
> >> instance. I only ever use SMTP connection so don't have to worry about
> >> direct queue injection by passing MailScanner.
> >> >> Drew
> >> >> Neil Robst said:
> >> >> > Yes... fingers crossed!
> >> >> >
> >> >> > Any other issues known with the 4.26-4 beta currently? What's the
> >> general feeling in the community of it's stability, etc?
> >> >> >
> >> >> > On Mon, 2004-01-19 at 22:07, Drew Marshall wrote:
> >> >> >> Just for my 2p, my server doesn't have a high load but I suffered
> >> duplicate mail. My old set up on Slackware didn't suffer, the new on
> >> Gentoo did :-(  . I'm not quite sure why but it seemed that the
> >> >> Postfix
> >> >> >> queue runner and MailScanner got in each others way with the
> >> result
> >> >> that
> >> >> >> MS picked up incomplete messages.
> >> >> >>
> >> >> >> <fingers crossed> Any way that's all in the past now </fingers
> >> >> crossed>
> >> >> >>
> >> >> >> Drew
> >> >> >>
> >> >> >> Neil Robst wrote:
> >> >> >>
> >> >> >> >Hi all,
> >> >> >> >
> >> >> >> >Just applied the 4.26-4 beta of MailScanner to my mail server,
> >> >> though
> >> >> >> I've
> >> >> >> >been unable to replicate the problem with the duplicate mails
> >> either
> >> >> >> before
> >> >> >> >or after (as expected) the upgrade. Do you know any details about
> >> that -whether it only manifested itself when there were lots of
> >> >> >> recepients
> >> >> >> >on the message or a high load on the server or what?
> >> >> >> >
> >> >> >> >Regards,
> >> >> >> >Neil
> >> >> >> >
> >> >> >> >
> >> >> >> >--
> >> >> >> >This message has been scanned for viruses and
> >> >> >> >dangerous content by MailScanner, and is
> >> >> >> >believed to be clean.
> >> >> >> >
> >> >> >> >
> >> >> >>
> >> >> >> --
> >> >> >> In line with our policy, this message has
> >> >> >> been scanned for viruses and dangerous
> >> >> >> content by MailScanner, and is believed to be clean.
> >> >> >> www.themarshalls.co.uk/policy
> >> >> >
> >> >> --
> >> >> In line with our policy, this message has
> >> >> been scanned for viruses and dangerous
> >> >> content by MailScanner, and is believed to be clean.
> >> >> www.themarshalls.co.uk/policy
> >> >
> >>
> >>
> >>
> >>
> >> --
> >> In line with our policy, this message has
> >> been scanned for viruses and dangerous
> >> content by MailScanner, and is believed to be clean.
> >> www.themarshalls.co.uk/policy
> >
>
>
>--
>In line with our policy, this message has
>been scanned for viruses and dangerous
>content by MailScanner, and is believed to be clean.
>www.themarshalls.co.uk/policy

--
Julian Field
www.MailScanner.info
MailScanner thanks transtec Computers for their support

PGP footprint: EE81 D763 3DB0 0BFD E1DC 7222 11F6 5947 1415 B654