Sent messages receiving delivery delay messages after CCR failover

We're having an issue where we are seeing delivery delay messages being sent to users who have sent messages that have already been delivered.  The catch is that these delivery delay messages are only sent when the primary node is patched, rebooted and the passive node takes over.

For example.  User A sends a message at 1:00 pm EST which is delivered at 1:05 PM EST.  Server A is the primary node and is patched and rebooted at 10:00 PM EST the same day.  Server B becomes the primary node ~10:05 PM EST.  At 10:09 PM EST a message is sent to User A stating the message he sent at 1:00 pm EST is delayed and will be retryed until XX:XX.  

Any ideas on this?  Our transport settings are at the default and I can't think of why a passive node being promoted to active would do this?  Only thing i can think of is the transport dumpster somehow impacts all this.
Who is Participating?
pgleekConnect With a Mentor Author Commented:
After working on this further i've been able to determine the reason.

The situation described above can occur when your mailbox cluster fails over and  encounters a lossy failover and at the same time some stores do not mount(CCR).  

When a lossy fail over occurs the promoted node will request that the transport dumpster on the Hubs resend the information it has in it's queue.  Since the transport dumpster only has delivered messages in it's queue, it is just making sure these messages got to their destination.  Since some of the mailbox stores did not come back online after the lossy failover the Hub sent a delivery delay message because the destination store was not reachable.

While the transport dumpster did seem to work as intended this seems almost flawed in that data can sit in that dumpster till it expires or new data over writes it and users can see delivery delay messages that are invalid and prompt calls to a Helpdesk, etc..
All Courses

From novice to tech pro — start learning today.