What steps will reproduce the problem?
1. Prepare a message to two recipients in different domains that live in
different STAs.
2. Ensure one of those STAs does not return a MDN
3. Dial down your mdnjobs.xml file to time out quickly, 1 minute. Then set
the clean up jobs to 1 day.
4. Send the message and wait for the MDN and a DSN for timeout.
5. Update the MDN table in the past. I choose 5 days in the past. Like so:
UPDATE [DirectConfig].[dbo].[Mdns]
SET [CreateDate] = DateAdd(day, -5, [CreateDate])
Watch the tables and your inbound message area. You will see a DSN generated.
But you have already generated a timed out DSN previously. This is the bug.
There are some indexes needed in the table to reduce contention when MDNs are
arriving and clean ups are running. This will also be included in this bug fix.
I have a solution but will take some time to get it through the builds.
Original issue reported on code.google.com by joeshook@gmail.com on 1 Oct 2014 at 12:35
Original issue reported on code.google.com by
joeshook@gmail.com
on 1 Oct 2014 at 12:35