Sending to servers with no MX record

In order to send email, the domain name (charlie.mydomain.com) needs to have either an MX record in the child domain table or needs to have a blank host record which points to an IP address.

A little ore simply, charlie.mydomain.com needs to be resolved to an IP address. IT either needs to be a hostname (an actual machine name) or if charlie.mydomain.com is a domain/subdomain (and not an actual machine), it needs an MX record or an A (host) record with a blank name (which will cause charlie.mydomain.com to resolve to that address).

You shouldn't attempt to reroute the delivery of the email to the parent domain.  Also, how could you configure Exchange to deliver email in the absense of MX records when you don't know where to reroute the email to?

Either the email address is wrong  or it's the recipient's problem.  They're saying "Hey, send me email!" and they're not set up to receive email.

I think I found the problem, well, part of it. I increased the timeout on the DNS forwards to give more time for the query. Still need to reroute to parent for some addresses(the parents do have MX records, just hope they have a better idea of where to send it, or access to an internal DNS). There also seems to be some difference in nslookup between NT and 2K. I get unresolved query's for mx records on NT, but resolves on 2K. Makes me worry though because 5.5 sits on NT...

Problem still occuring, only way around it seems to be to configure message delivery for the problem domains to route to their parent domain. Thinking it might be an exchange name resolution configuration. DNS is doing recursive queies, exchange seems to have some DNS problems that seem independent of the server (server resolves, exchange doesn't). IMC is configured to use DNS for resolution, except for these problem domains. Got this message when I tried to use DNS solely "User unknown to Lyris ListManager". Both source and destination server are exchange, seems like message was routed to a Lyris server somehow. Am I missing something?

You say that the server resolves DNS but Exchange doesn't.  Could you explain how you arrived at that conclusion and how?  It doesn't seem to make sense to me.

Also please provide an example of one of the child domain names.

If I do an nslookup to els.wylelabs.com at the exchange server, and set type to MX, I get:

wylelabs.com
        primary name server = ns1.pbi.net
        responsible mail addr = postmaster.pbi.net
        serial  = 200206130
        refresh = 3600 (1 hour)
        retry   = 900 (15 mins)
        expire  = 604800 (7 days)
        default TTL = 7200 (2 hours)

which I assume relays to the mail servers on the DNS hosting the parent domain and their MX records. The IMC queue however will not forward to the mail and return a "network host resolution error". I get arround this by forwarding to the parent domain (IMC static entry). That "Lyris ListManager" error is a new one, don't know what that's about. Thought the problem was isolated to this one domain, but has appeared for several this week. All child domains without explicit MX records.

It's an incorrect assumption that it relays to the parent domain.  It doesn't.

There is no MX record because els.wylelabs.com is a host name [216.100.104.186], exchange will attempt to delive the mail there on port 25. And that host IS running Microsoft SMTP.

Your Lyris error message is because you're forcing mail to the parent domain's mail server and the parent's mail server is rejecting it because it doesn't belong to it.  So the Lyris stuff is not a REAL problem, it's a result of your redirection efforts.

Can you resolve els.wylelabs.com as a host name?  Can you open an SMTP connection to it?

Yeah, I caught that listserve.wylelabs.com a minute ago. I can resolve els.wylelabs.com at the server, I can telnet to port 25 from the server, but the mail won't leave the queue. Is there a setting somewhere that say's "don't attempt to deliver if the host doesn't have an MX record". I know the redirection is not desireable, but it's been working.

Infact, the list server error message appeared when I removed the redirection entry yesterday and relied on DNS only. I suspect it will happen again sometime because the costs are all the same.

I just sent a bogus email to foobar@els.waylelabs.com and it was accepted and shortly returned a recipient failure message.  So my email left my Exchange server queue.

And no, there isn't a "don't deliver if no MX record" setting, not that I know of anyways.

Did you do your name resolution tests on the Exchange server itself?  Do you know that it can resolve this host name?

Yep, right at the same console. Comes back the host record:

Non-authoritative answer:
Name:    els.wylelabs.com
Address:  216.100.104.186

I think it was working from DNS but that listserver box being the same cost has a one in seven chance of getting mail routed to it, and is not aware of any exchange boxes. I'm still tring to understand if this is a problem with my server, of a result of their DNS administration. Either way, it's my job to make it stop. If it helps, the DNS we forward queries to is dfwns2.airband.net.

should this els record be an "MX' record? Does it have to be?

The domain name "els.wylelabs.com" can either be a host name (which this is) or a domain name.  If it's a domain name, the domain needs to have an MX record pointing to a real IP address or a blank host (A) record pointing to a real IP address.

"that listserver box being the same cost", what do you mean by this?

Do you still have a statis rule which redirects the els.wylelabs.com address?  If so, where do you redirect it to?

By cost, I mean preference, from the return below for Wylelabs.com (MX), listserver has the same preference as the most of the others. Therefore I would think it has a one in seven chance of being directed mail for wylelabs.com. Not sure what you mean by statis rule, the only rule I set up was on the IMC under message delivery, an exception to deliver by DNS was to relay mail for els.wylelabs.com to wylelabs.com.  I removed this a few days ago and restarted the service so DNS should have been the only delivery method. That's when I recieved the listserver error. That got some visibility so I set up the rule again, but this time pointed directly to hou.wylelabs.com. I didn't get any bounces, so I'm hoping it's relaying it to the right place. I'm going to award you the points, but I'm still trying to pinpoint the problem, why do you think it was relayed to the listserver box after the rule was removed and service restarted? Are there any places to specify an alternate DNS server within Exchange that may need to be changed, or even a timeout setting? I'm thinking exchange may be taking a longer time to resolve the name than from what I can see by doing nslookup on the mail server. BTW, my mail server is pointeing to an internal DNS server (NT4)that's configured to user forwardes and operate as a slave server. It tforward timeout has been increased to 15 seconds and forwards to an my external DNS (W2K) that's configured to forward querys to my ISP, timeout of 10 seconds using recursion. That's everything, I think, any ideas?

wylelabs.com    MX preference = 100, mail exchanger = smtp-relay.pbi.net
wylelabs.com    MX preference = 10, mail exchanger = arl.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = hmp.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = hnt.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = hou.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = nor.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = mail2.wylelabs.com
wylelabs.com    MX preference = 10, mail exchanger = listserv.wylelabs.com

yeah, I guess I can do that. I think I'm going to remove the rule, reboot this weekend, and just hope it stops happening. This isn't the only domain with the problem, I was just hoping there was something else I'm missing. Thanks for the help!

seems like my DNS had a problem resolving els a few minutes ago. I tried to restart the IMC and it wouldn't because it couldn't fine the hou.wylelabs.com server. Took out that entry, set my internal server to not be a slave server and added my ISP's servers to it's server list, then pointer my email server to the external DNS. Had to hammer nslookup for a minute before I got an answar for els again.....

Does anyone know how to increase the DNS timeout?

Ian