Alessandro Vesely writes:

> Hi all,
>
> I apparently lost a message.  I found it in the log:
>
> Jun 11 20:57:34 north courierd: Waiting.  shutdown time=none, wakeup  
> time=none, queuedelivering=1, inprogress=1
> Jun 11 20:57:34 north courierlocal: id=00000000005DC03F.
> 000000004FD63F99.00003F1E,from=<xxxx@yyyy...>,addr=<zzzz@tana...>,size=5061,success:  
> Message delivered.
> Jun 11 20:57:34 north courierd: completed,id=00000000005DC03F.
> 000000004FD63F99.00003F1E
> Jun 11 20:57:34 north courierd: Waiting.  shutdown time=Mon Jun 11 21:26:22  
> 2012, wakeup time=Mon Jun 11 21:26:22 2012, queuedelivering=0, inprogress=0
>
> But it wasn't in the folder it should have gone to.  It isn't
> anywhere.  I only have seven (other) messages from that guy,
> so I run a
>    find ./Maildir -type f | xargs grep -li xxxx@yyyy
> search, but it only found the same seven messages as the MUA.
>
> It is the second time this happens to me, in all these years,
> but the previous time I hadn't been able to locate the log
> entries so I gave up.  How can this be at all possible?
>
> The Maildrop recipe consists of if (...) { ...; to "./..."; }
> and terminates with to "./Maildir/".  A disk error should have
> been reported prominently, shouldn't it?  (I run a Debian
> squeeze.)

Well, you have an alternative delivery there, so that's maybe where it went.

Otherwise, there are only two possibilities. It was delivered, but  
subsequently to that some other process deleted. Or, there was a delivery  
error, but Courier thought that the message was accepted for delivery.

Double-check how you have Courier invoke maildrop; using either  
DEFAULTDELIVERY or from a .courier file; that it's a simple invocation. If  
you have a whole shell mini-script in there, that invokes maildrop,  
depending on how it's done maildrop's exit code may get lost, and Courier  
has no indication that mail delivery failed.