I`ve done it twice, with install and reinstall Edge server 2013. 3d time, i`ve used clean server 2012 installation (not a template) and Exchange installation.

We have:

LAN Exchange 2013 organization, 2 servers (CAS + MBX x 2) SR1 and SR2, placed in DAG,  SR1 takes 25 inbound port through CISCO ASA and its all just fine. I installed  EDGE server in DMZ, 2nd NIC placed in "LAN-Servers". ACL between LAN-DMZ "ANY-ANY".

1. Configure Internal SMTP server on Transport Configuration Set-TransportConfig  -InternalSMTPServers 10.1.20.31, 10.1.20.32
2. On the Edge, add DNS-suffix  "domain.local", edit hosts file and add LAN Exchange servers
3. ADD local DNS Host (A)  EDGE1.domain.local with LAN IP 
4. LAN NIC leave w\o geteway, add local DNS servers 
5. DMZ NIC with gateway, DNS LAN servers either (none of external DNS records)
6. From Edge made New-EdgeSubscription -FileName "C:\Edge.xml" -Force
7. From mailbox New-EdgeSubscription -FileData ([byte[]]$(Get-Content -Path C:\Edge.xml -Encoding Byte -ReadCount 0)) -Site msk-site 
8. From mailbox  Start-EdgeSynchronization -Server SR1 -TargetServer Edge1 -ForceFullSync, success.
9. Turn off manual created External Send Connector, and checked outbound mail flow, success.
10. From mailbox SR1 Set-SendConnector EdgeSync Inbound to Default-First* -Port 2525 (i`ve tryed 25 either)
11. Remaped inbound 25 port from CISCO ASA to the EDGE server DMZ Nic. 
12. Checked telnet 25 with External MX record, passed test user email availability with ms connectivity analyzer .

And... there is no inbound messages anymore. Get-Queue on the Edge, shows me inbound queue to internal Exchange servers with next error:

NextHopDomain                    :
sr1.domain.local,sr2.domain.local

Status                           : Retry

MessageCount                     : 31

LastError                        : [{LRT=16.07.2014 0:31:47};{LED=441 4.4.1 Error encountered while communicating with

                                   primary target IP address: "Failed to connect. Winsock error code: 10060, Win32 erro

                                   r code: 10060." Attempted failover to alternate host, but that did not succeed. Eith

                                   er there are no alternate hosts, or delivery failed to all alternate hosts. The last

                                    endpoint attempted was 10.1.20.25:25};{FQDN=sr1.domain.local,sr2.domain.local};{IP=10.1.20.25}]

We have no IP 10.1.20.25 and we`ve never had. Local IP`s is  SR1 -10.1.20.31, SR2 - 10.1.20.32. IP 10.1.20.25 cant be resolved, and has no ping, there is no server parameters with this IP (connectors configuration). From Edge server getting true answer by cmdlet:

Get-TransportConfig | fl InternalSMTPServers

InternalSMTPServers: 10.1.20.31, 10.1.20.32 

Well, im done with it. No ideas at all.

Last things that i have:

1. Add directly IP adresses in the "EdgeSync - Inbound" connector with replacing "--" type of smarthosts
2. Turn ESMTP off on the CISCO ASA firewall. 

• Edited by Oleg.A Thursday, July 17, 2014 9:48 AM

Success!

I perfomed the first thing - "Add directly IP adresses in the "EdgeSync - Inbound" connector with replacing "--"", and all goes work fine!

Can somebody explain, why was that happened?

• Marked as answer by Oleg.A Monday, July 21, 2014 4:25 AM

Success!

I perfomed the first thing - "Add directly IP adresses in the "EdgeSync - Inbound" connector with replacing "--"", and all goes work fine!

Can somebody explain, why was that happened?

You sir are a star.  I've searched so many threads and forums trying to find this answer.  Thank you so much!

In my scenario, I had the main Exchange server working fine with email going from eh internet straight to the CAS/MBX server, and outbound via the ET.  Everything worked fine.  Then, when I asked the mail to be delivered to the ET to then be passed onto the MBX, it completely messed up.  Why they ever allowed the '--' setting to be introduced as a default config is a mystery!  Thsi clearly doesn;t work so is either very poor testing from MS, or they simply want to make this process as difficult as possible so we all jump into Office365.  They can shove that up their backsides!

Thank you MS for wasting countless hours of my life!

Also, if anyone has issues sending out after installing an ET,

• This was purely a DNS issue.  Again, when the ET is built and added, the TransportService ExternalDNSAdapter and InternalDNSAdapter were both set to {}.  After setting these in the EMS with the Set-TransportService {server name} internalDNSadapter  {ip of DNS} outboundDNSAdapter {ip of DNS} this worked fine.

I can now go to bed......no gym, no lunch breaks.....I'm a grumpy human! :(