One of my customers has his Domino 7.0.4FP1 server pointed to messagelabs for outbound mail routing. He gets a lot of transient routing failures (not always the same host):
Router: No messages transferred to CLUSTER8OUT.EU.MESSAGELABS.COM (host mail195.MESSAGE) via SMTP: The remote server is not a known TCP/IP host.
Note that the host name in the error report is truncated. Digging and diagnosis tells us that in the in the DNS, that domain has MX records for 11 hosts, all with a preference of 10. Which means that Domino picks one of these at random - or, more probably, cycles round the list. But it appears that the list that it stores gets truncated somewhere, so every so often - about 10% of the time, in fact, the message route fails and it tries again with another server.
This annoys my customer - he dislikes seeing errors in his server log that he can't resolve - and doesn't want to ignore it or just to filter it out. Anyone seen this one before. Got any ideas?
Here's what I get from an MX query to dsn4.messagelabs.com for cluster8out.eu.messagelabs.com.
> cluster8.eu.messagelabs.com
Server: dns4.messagelabs.com
Address: 216.82.242.229
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail205.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail206.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail178.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail179.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail186.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail188.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail189.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail193.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail194.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail195.messagelabs.com
cluster8.eu.messagelabs.com MX preference = 10, mail exchanger = mail204.messagelabs.com
>
Mick Moignard November 13th, 2009 02:59:50 AM
Try these debugs to get more info on what is going on:
SMTPDebugControls=1
Debug_TCP_Resolver=1
Also make sure you do not block DNS queries on TCP allowing only UDP.
{ Link }
I've seen something similar to this before when using MessageLabs.
In that particular case, the clients outbound SMTP was inadvertently pointing to a Messagelabs 'inbound' gateway, rather than a dedicated outbound one.
It may be worth checking there isn't a dedicated 'outbound' SMTP gateway that you should be using instead.
I get the same on customer's pair of R7.0.3 servers... originally thought to be communication issue as they had problems with their internet connection but proved to be the same.
An R8.x server seems OK with cluster8out.eu.messagelabs.com so have just added specific hosts from the messagelabs cluster as the relay server for now.
If anyone has come up with a fix for R7 servers and this issue please let me know!
Steve
Any solution to this? We are switching over to MessageLabs and are experiencing the exact same issue.