Exchange server sends mail fine right after a reboot

[NetworkDOC]

Once I reboot the exchange server mail flows great. After a while though (a few days sometimes) it stops. I rebot and all mail flows fine. It only hangs on a few domains too.

I did the correction to fix the EDNS 0 issue. That allowed mail to flow out to hotmail/yahoo users fine. This is a hmmmmmmmmm Problem...

Any thoughts?

 

[NetworkGuy101]

Whats in your event logs? Is there any exchange related errors around the time mail stops sending?

 

[NetworkDOC]

I double checked (recreated the problem). NO event logged.

No errors report at all. I do get the automated warning, then eventually a failed report.

This is an automatically generated Delivery Status Notification.

THIS IS A WARNING MESSAGE ONLY.

YOU DO NOT NEED TO RESEND YOUR MESSAGE.

Delivery to the following recipients has been delayed.
user@theirdomain.COM

Yet if I let it go (without rebooting the server) I get:

Your message did not reach some or all of the intended recipients.

Subject: URGENT: Cannot Send Mail
Sent: 12/29/2004 8:06 PM

The following recipient(s) could not be reached:

The user(me) on 12/31/2004 8:11 PM
Could not deliver the message in the time limit specified. Please retry or contact your administrator.
<Exchange.theirinsidedomain.local #4.4.7>

If I reboot the server, all mail flows fine.

MIGHTY Strange.....

 

[NetworkGuy101]

Just a couple of ideas to ensure DNS is working properly:

Try an NSlookup from your Mail Server to ensure you can lookup up other mx records.

Command Prompt:

NSLookup

set type=mx
yahoo.com

You should get back a list of MX Records

Also you can go to

http://www.dnsstuff.com

and do an nslookup of yourself to see if your mx records are listed properly. If your SMTP Connector is setup correctly it wouuld be good to get DNS out of the way as a possible problem.

 

[NetworkDOC]

DNS is resolving fine. All I have to do is reboot the exchange server every two days and mail flows to everyone fine. If I stop some domains will not flow. The sender receives the undeliverable message above.

I can lookup my domain and other domains mx records without issues.

When I run a DNSreport it comes up fairly clean no red a few warnings (soa timer etc). Nothing bad. Same thing when I run the same test against mydomain.com

ODD.... Simply ODD......

Any other Ideas? I will try them out.....

 

[NetworkGuy101]

Instead of rebooting have you tried just stopping and restarting your SMTP Virtual Server? Also, do you have a reverse record for your server in your external DNS? What kind of firewall are you behind? There are some issues if you are behind a pix.

 

[NetworkGuy101]

Also if you in Exchange Manager, if you go to your mail group then protocols--SMTP--Right click virtual server then select delivery and click advanced, Do you have any DNS servers listed here?

 

[NetworkGuy101]

Do you have Symantec on this Server?

 

[NetworkDOC]

I have tried restarting the smtp virtual server, then tried restarting the exchange services, then tried restarting the smtp service. None of these did it. I tried stopping the symantec mail service, the av service didn;t do it either.

In my SMTP virtual server I do have two DNS servers listed there. This appears to be working ok as mail flows most of the time normally.

I am behind a pix. I did the DNS fix, or should I say no dns fixup 512 (something like that). I also did the EDNS0 fix. dnscmd ServerName /Config /EnableEDnsProbes 0


Great suggestions.... Any more?

 

[Chambers]

Maybe it's just your NIC going bad???

Also, I know MS says to set the SMTP Server to a actual IP address and not "all unsigned" or whatever that option is. I'm having a similar issue where MAIL just gets piled up after a few days and I need to reboot. I am going to change the NIC this weekend to see if it alleviates the problem. Fingers crossed!

 

[kzn]

I know this sounds silly but how much drive space do you have on the system drive? exchange needs more than 200MB on the system drive. I have seen a server that had a similar problem, this was due to insufficient drive space.

 

[NetworkDOC]

The server has 50+Gb Of useable space (about 90% of the space is free). Probably not that....

I did designate the nic (this am) to dee if this indeed matters. I also told the ther NIC (there are two not to register in DNS. We'll see if this matters....

Tks

 

[NetworkGuy101]

The is a KB out there for mail not flowing properly behind a PIX. I had the same issue. I will post it when I can find it or If I remeber what I did will post that also. I think you just have to do No fixup SMTP though. I believe the smtp fix up has problems with Microsoft servers. Ill get back with the info.

 

[NetworkGuy101]

Here is the link to the KB Article. Also you will see links in the Doc to Cisco.

http://support.microsoft.com/?scid=.../servicedesks/bin/kbsearch.asp?article=320027

 

[NetworkDOC]

I did the pix, No DNS Fixup command. It allowed mail to flow, but then every few days it will stop sending out mail.

 

[NetworkGuy101]

What about the No SMTP Fixup?

 

[NetworkGuy101]

On my Pix I still have the DNS Fixup but I have:

No Fixup protocol SMTP
In My outside-->IN ACL I am allowing SMTP to my mail server and IDENT.

I had the same problem. Receiving mail was flaky from yahoo or hotmail but as soon as I added the No Fixup SMTP mail started coming right in. I am running Exchange 2003 Server and Pix 6.1

 

[Chambers]

What was the solution to this?

 

[NetworkDOC]

The verdict is still out.

I had to reboot Friday to get mail flowing again. I am about ready to call MS to see if they can assist me in getting a resolution. I have about exhausted any ideas that I have been given and have thought of on my own.

Another thought... I hadn't rebooted the server since I had put the no smtp fixup command on the pix.

I guess I can wait a few more days and test to see if it indeed fixes it.

I am scheduled to go back on site there Wednesday.

I intend to post a resolution once I am confident the problem is resolved. Someone else may run into this....

 

[Chambers]

Well just for the record I had a similar issue and all I did was replace the NIC in the server (why not huh?) and it's been working fine since then. Perhaps that's your issue also? Try it, it JUST may work