Re: [Exim] backup delivery (was: ACL verify=sender)

Pàgina inicial
Delete this message
Reply to this message
Autor: Tony Finch
Data:  
A: Kjetil Torgrim Homme
CC: Tony Finch, David Woodhouse, tsh, exim-users
Assumpte: Re: [Exim] backup delivery (was: ACL verify=sender)
On Mon, Jan 05, 2004 at 08:30:49PM +0100, Kjetil Torgrim Homme wrote:
>
> hmm. losing the queue is annoying, but it should typically not amount
> to more than a few seconds's worth of e-mail, i.e. on the order of a few
> dozen messages.


We have thousands of messages on the queues on these machines, the vast
majority being frozen double bounces and many of the rest being bounces.
These should go away when I get a callout setup I am happy with.

A future problem that we need to protect against is messages that are
queued because of users' quota limits on the Cyrus servers. This is
the main reason for engineering a careful solution.

> to protect against a failing Cyrus server, we have set up a normal
> unseen shadow delivery. it delivers to
> file = MAILDROP/${substr_0_10:$tod_log}/$local_part
> one file per user maintains the order of delivery, which is nice.


However it doesn't take into account Cyrus's single instance store.
Cyrus will re-merge the messages, but it makes re-injecting them
less efficient.

> we can also easily help Eudora users who pressed the wrong button and
> deleted all their e-mail.


We have patches to Cyrus which make it delay expunges and folder
deletions so that we can recover messages for users. Instead of
immediately removing messages from disk there's a separate expire
job (like on a news server). I'll be talking about this at the
UKUUG Winter Conference next month. There's some information at
http://www-uxsup.csx.cam.ac.uk/~dpc22/cyrus/

Tony.
--
f.a.n.finch <dot@???> http://dotat.at/
LANDS END TO ST DAVIDS HEAD INCLUDING THE BRISTOL CHANNEL: WEST 5 BACKING
SOUTH AND DECREASING 3 OR 4 LOCALLY 5, LATER INCREASING 4 OR 5 LOCALLY 6. RAIN
SOON CLEARING, BUT WITH PATCHY DRIZZLE LATER. MODERATE OR POOR IMPROVING GOOD,
BUT DECREASING MODERATE LATER. ROUGH DECAYING MODERATE TO ROUGH.