Exception: An existing connection was forcibly closed by the remote host
Question asked by Steve Vibert - November 16, 2015 at 4:03 PM
Unanswered
Since upgrading to 14.4 I'm seeing issues with two users who are forwarding mail from their corporate email accounts to personal accounts.  Both users are forwarding to different domains but the errors in the delivery log are the same:
 
09:36:57 [22850] Exception: An existing connection was forcibly closed by the remote host
   at RelayServer.Clients.SMTP.ClientConnectionSync.SendLine(String line)
   at RelayServer.Clients.SMTP.SmtpClientSession.SendLine(String #a8)
   at RelayServer.Clients.SMTP.SmtpClientSession.SendData(StreamReader #bJ)
   at RelayServer.Clients.SMTP.SmtpClientSession.SendData()
   at RelayServer.Clients.SMTP.SmtpClientSession.#OI()
   at RelayServer.Clients.SMTP.SmtpClientSession.StartProcessing()
 
When a forwarded message like this fails, SM attempts to resend the message every second.  I'm not sure if there's a correlation but after 6+ hours of resend attempts, the mail server slows to a grinding halt even though the CPU and memory utilization appears to be normal. 
 
It was so bad this afternoon that I couldn't log into webmail (constant progress spinner on the SM GUI).  I used a remote desktop session to access the server and it was dog slow.  Once conneted remotely I was able to get into webmail and I eventually got to the "Waiting to Deliver" spool page. Three messages had been attempting to deliver to these same 2 recipients since approx. 9AM.
 
Restarting the Application Pool and website had no effect and it literally took minutes for the server to responds to mouse clicks.  I eventually rebooted the server afterwhich everything was back to normal.
 
This is the second time in 5 days that this has occurred--the first time was the day after upgrading. Again, I'm not sure if there's a correlation but it seems pretty coincidental.
 
Anyone else seeing similar problems?
 
Thanks -- Steve

2 Replies

Reply to Thread
0
Steve Vibert Replied
I'm now seeing this on non-forwarded outgoing mail. I've got 2 messages that have been sitting in the waiting to deliver spool. Both are from the same sender. One message has been in the spool for an hour and 10 minutes and has had 9 delivery attempts. The other for an hour and 2 minutes and has had 6 delivery attempts.

The weird part is that one of the message was received by the recipient. I'm guessing that the second has as well but I can't confirm that for sure.

Restarting the SM server, or Application Pool has no effect.
0
Steve Vibert Replied
It occurred to me this morning that I've seen this before.  I looked through my notes and the last time we had this issue was in July of 2014 and it resulted in us being blacklisted on Microsoft's email domains.  Back then, the fix was to disable ESMTP inspection on our Cisco ASA. 
 
I had a look at our ASA config this morning and lo and behold ESMTP Inspection was turned on again.  We made some significant network changes over the summer that involved some pretty major ASA configuration changes and I'm guessing we omitted turning it off.
 
The nature of the failure was very random and infrequent back then which might explain why it took so long to manifest.  However, it's never caused the server unresponsiveness issues I saw yesterday.
 
Anyway, I'll keep an eye on the logs and will update this thread when I know more.

Reply to Thread