Bug? Bandwidth for failed messages.

lbauer

Verified User
Joined
Dec 26, 2004
Messages
16
I think there may be a bug or omission in the bandwidth calculations. When a customer tries to send a message that is too large, and the message is rejected, the bytes that were sent are not included in the bandwidth calculation. Here is a sample error message from /var/log/exim/mainlog:

2006-05-20 13:11:45 1FhUw3-0004df-7V rejected from <[email protected]> H=xx.yy.net ([192.168.3.69]) [xx.xx.xx.xx]: message too big: read=37738252 max=20971520

Names have been changed to protect the innocent. The first [email protected] is the sending users email address. xx.yy.net is the reversed actual ip address. xx.xx.xx.xx is the actual ip address.

Background: This became an issue today as I noticed that bandwidth usage was way above normal. I looked at the customer usage and it didn't show anything out of the ordinary. I put a traffic monitor on the DirectAdmin box (iftop - great program) and it showed it was inbound smtp traffic from a particular ip. I greped the mainlog for that ip and found a customer had a 37mB file in their outbox and it was getting to resent every 5 min. Average utilization was a complete T1 (1.5mbps). Exim would reject it once they had uploaded a large portion of it. 5 min later it would start again.

It would seem to me that DirectAdmin should include the amount of traffic sent before the rejection in the bandwidth tally.
 
Back
Top