Had an issue with a couple users trying to send to 1 specific domain that just seemed to have stopped working. Exchange queue on our Edge server was showing the following error: 4.4.2 Connection Dropped
After several hours of research, and determining it was nothing on our side, I contacted the other sides IT team who reached out to rackspace who was hosting their email.
Logs on Edge on our side showed the following
C:Program FilesMicrosoftExchange ServerTransportRolesLogsConnectivity
2016-07-01T00:23:29.590Z,08D387F7316D769B,SMTP,domain.com,>,Established connection to XXX.XXX.XXX.XX1
2016-07-01T00:23:30.043Z,08D387F7316D769B,SMTP,domain.com,>,Established connection to XXX.XXX.XXX.XX2
2016-07-01T00:23:51.200Z,08D387F7316D769B,SMTP,domain.com,>,Failed connection to XXX.XXX.XXX.XX3 (0000274C)
Why was it establishing a connection, but moving onto the next MX record still?
Ran MX toolbox
Ran Microsoft Exchange Connectivity Analyzer which showed it wasnt able to connect to the 3rd MX record
Questioned what the 3rd MX record was, and it was a stale on-prem email record that was never removed. They updated DNS to remove this. Let it propagate the internet
Ran a ipconfig /flushdns on Edge server
Cleared the messages on the Edge server in this domains queue, Suspended the Queue, Resumed it to clear it out, sent another email and it went through successfully.
Cause: Stale MX record on recipients side
Conclusion: Only explanation I can come up with is that the first 2 MX records for rackspace were unavailable and it tried and hung onto the 3rd stale MX record and kept trying to use that.