[exim] Immediate bounce after mail is received

Hello all,

On one of the systems I administer, I have a strange problem I can't
really get a grasp of why it happens.

Maybe one of you can help me out here.

It started when we noticed the following block in the exim log in the
server, which act as an smtp outbound server on the infrastructure:

(I anonymized the email address it goes to, but I tried to leave as
much information as possible intact - I just don't have the liberty to
write it non-anonymized)

[root@lilzmailso01 exim]# exigrep bw@??? *
2008-10-06 10:07:02 1Kml7N-0006mb-SY <= ha@???
H=cpe90-146-29-228.liwest.at (auerad1a8f42f7) [90.146.29.228] P=esmtpa
A=fixed_login:uid=hklvappx S=1822
id=D3BDF142D43C448CB7D782EEFA768877@auerad1a8f42f7
2008-10-06 10:07:07 1Kml7N-0006mb-SY == bw@??? R=dnslookup
T=remote_smtp defer (-44): SMTP error from remote mail server after
RCPT TO:<bw@???>: host mail4.internic.at [212.126.66.30]: 450
<bw@???>... MAILBOX temporarily unavailable(GREY_SENDER:
ha@???)
2008-10-06 10:07:07 1Kml7N-0006mb-SY ** bw@???: retry timeout exceeded
2008-10-06 10:07:08 1Kml7N-0006mb-SY Completed


[root@lilzmailso01 exim]# exigrep 1Kml7N-0006mb-SY
(This yields the same block as above, and additionally the following)
2008-10-06 10:07:08 1Kml7T-0006nl-Or <= <> R=1Kml7N-0006mb-SY U=exim
P=local S=2891
2008-10-06 10:07:08 1Kml7T-0006nl-Or => ha@??? R=remotes
T=remote_smtp H=192.168.224.90 [192.168.224.90] X=TLSv1:AES256-SHA:256
2008-10-06 10:07:08 1Kml7T-0006nl-Or Completed

[root@lilzmailso01 exim]# exinext bw@???
Route: bw@???:<ha@???> error -44: SMTP error from remote mail
server after RCPT TO:<bw@???>: host mail4.internic.at
[212.126.66.30]: 450 <bw@???>... MAI
first failed: 30-Sep-2008 09:53:25
last tried: 06-Oct-2008 10:07:07
next try at: 06-Oct-2008 16:07:07
past final cutoff time

[root@lilzmailso01 exim]# exim -brt bw@???
Retry rule: * * F,6h,15m; G,16h,1h,1.5; F,4d,6h;

The problem I see here is, that there were successful deliveries
between Sept. 30th and Oct. 6th, so I thought that the retry data
would be deleted after a successful delivery.

And there are much too few delivery attempts in the time for the retry
configuration to justify a hard bounce / message freeze for the first
message anyway.

Can anyone provide some hints what and why it is happening here? Do
you need some more information like configuration of the Exim? Or is
there any tool I could run to gather some more information about this?

The Exim running here is a 4.66 on RHEL 4.

Thanks in advance for you time and help!

Jens