Link to home
Start Free TrialLog in
Avatar of Todd Gerbert
Todd GerbertFlag for United States of America

asked on

Outbound SMTP Messages Stuck for Certain Domains

I'm running a single Exchange 2003 system, two virtual SMTP servers bound to different IP addresses (one accepts anon connections but only from upstream server, the other accepts from any system with auth reqd).  Both virtual servers are configured to deliver direct, DNS servers are NOT listed in SMTP server config - DNS settings are configured properly on bound network adapter, external A and PTR records exist for both virtual servers.  Server is behind a PIX running 6.3 with "no fixup protocol smtp"

I am getting outbound messages for certain domains, like aol.com, comcast.net, and a couple others, that just don't seem to go anywhere - other domains don't seem to have any problem.  On my server, I can run nslookup to get MX records for affected domains and successfully deliver messages with telnet.  I've tried re-applying Exchange SP2 to no avail.
Avatar of Nuno Martins
Nuno Martins
Flag of Portugal image

Avatar of Todd Gerbert

ASKER

Thanks for the suggestions, but neither of these seem to apply.

I should also note that I am running McAfee Group Shield and VirusScan (I've excluded Exchange directory tree et al from OnAccess virus scanning) and Exclaimer (which appends a disclaimer to outbound messages).  I've also tried removing these without any luck - it doesn't seem to have an impact one way or the other.  I also had temporarily setup a smart host, just for testing's sake, and mail flow was fine.
Considering the domains you have listed, the problem could be down to your DNS configuration. I would check that is correct.
Can you telnet to the MX records of those domains?

Simon.
My DNS servers are okay, and I've tried using Level 3's DNS servers as well.  The DNS server addresses are listed on the network adapter's TCP/IP config, NOT on the Virtual SMTP Server config.

I can telnet to the MX records of the affected domains on port 25, and can manually deliver an e-mail using telnet.
I wasn't referring to DNS servers.
I was referring to DNS, as in how your server is seen by the Internet. Ensuring that MX records, DNS and reverse DNS records are all valid.

Simon.
MX record points to 3rd party filtering service (Postini), A and PTR records are correctly configured, and match the name on the Virtual SMTP server "Delivery" tab (which also matches the netbios name).
When the messages are in the queues what error does it say?

Simon.
Usually it reads "No additional information is available", sometimes "connection dropped by remote host"
ASKER CERTIFIED SOLUTION
Avatar of Sembee
Sembee
Flag of United Kingdom of Great Britain and Northern Ireland image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I've temporarily re-routed outbound mail to another smart host on my network - just another system running IIS's SMTP server.

I don't think it's DNS and I don't think the connection is really being dropped...could it be the IIS metabase?
If the IIS metabase was damaged then it would be happening to all domains. Plus the email still has to go through SMTP to be sent to your smart host. SMTP logging may tell you whether it is connecting and being dropped.

Simon.
I will crank up the logging to max and remove the smart host so I can get some error info...
Why have you asked for the question to be deleted?
You have been provided with significant amounts of information and the last post stated that you were waiting to see what the logs stated.

Simon.
Because I know mail services will be interrupted in order to get any error messages out of the logs, and I haven't had time to do so during off-hours, and probably won't for a few weeks because of other projects - I intend to re-visit this in a month or so.
Not really a good reason to ask for the question to be deleted. If you are unable to look at the issue at the moment then leave the question open. There are questions still open from this time last year.

Simon.
I intend to...leave the question open.
I've removed anti-virus and disclaimer software, re-applied SP2, and re-installed disclaimer & anti-virus.

Thus far my test messages to a few different domains seem to be okay.  I left diagnostics logging at maximum, and I'll see how things go during the week after while the server is under a more typical load.
I'm still getting messages stuck in queues for most domains (interestingly, some domains never get stuck - like hotmail).

Right now ESM shows about 50 messages in 3 queues, and "no additional information available."

Event viewer shows 1 occurence of event id 4000 "The remote server did not respond to a connection attempt" (this is for yahoo.com), and 6 occurences of event id 4006 "The remote host dropped the connection" (for yahoo.com, gmail.com, verizon.net, enter.net, and a couple others).

System event log shows 2 occurences of DCOM event id 10009 "Unable to communicate with any configured protocols", 1 for yahoo.com and 1 for enter.net.

On the stuck messages I can run an MX nslookup on the exchange server, telnet to the smtp server and deliver mail to the recipient of the stuck message.

I'm wondering if it's not a performance/resource availability issue...?
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
After reviewing the logs as you suggest, I've come to the conclusion the server is is hanging while attempting to make outbound connections (disk queue length in particular is high) - some remote SMTP servers timeout slower than others, hence some domains going and others not.  My has been to replace the server - so now I'm muddling my way through Exchange 2007. ;)

Thanks for your help, Simon!