Fog Creek Software
Discussion Board

Mail server timing out...

I've been having issues with large e-mail, and FB getting into an infinite loop of trying to load an e-mail message but not deleting it on the mail server (thus it continues to try to reload the message every time it poles the account).

Here's a potion of my dispatcho.txt file where I'm was testing the problem.

Reading user,pass,server,port,url
Setting normal (non-secure) mode
Attempting to log in
Num Messages...
Publish success
Com Error in CheckMail: An established connection was aborted by the software in your host machine, possibly due to a data transmission timeout or protocol error. PowerTcp will close the connection when this error is encountered.
SOFTWARE\Fog Creek Software\FogBUGZ\[PATH_TO_APPLICATION]/FogBUGZ/website\dispatcho

This seems to indicate to me that the mail server is timing out, and when dispacho tries to go back to delete the message i just processed, it gets an error.  (My test message normally takes around 5 minutes for dispatch to process and load into the DB)

I've done some investigation on this, and it doesn't seem to be anything related to the "Kiss of Death" scenario for pop servers. I'm certain there are no other pop clients polling the account in question (I even created a brand new account just to test this out).

Additionally 5 minutes is much shorter than the 10 minutes that RFC 1939 says that a server must wait before timing out the connection.

Am I chasing the right demon (deamon :-) ) here, by suspecting my mail server being the cluplirt, or is there something else I can change on the FB/Windows side that will affect this?

Sean Ennis
Thursday, June 24, 2004

I think you've pinpointed the exact problem successfully.

Dispatcho takes a long time to publish the message and hence, when it goes back to delete the message the POP connection has been closed, so it can't.  Then it reconnects to the server and starts all over again.

Can you turn on logging for your pop server?  Which server are you using?

Michael H. Pryor
Fog Creek Software
Thursday, June 24, 2004

I've running my pop server on RedHat 9 usiung the default server ipop3d (I'm told that this is WU-IMAP).

Here's what I et from the servers:
# ipop3d
+OK POP3 mailserver1 v2001.78rh server ready

# imapd
* OK [CAPABILITY IMAP4REV1 LOGIN-REFERRALS STARTTLS AUTH=LOGIN] mailserver1 IMAP4rev1 2001.315rh at Wed, 23 Jun 2004 11:16:32 -0500 (CDT)

I'll see what I can find in the log files. What appears in /var/log/maillog doesn't appear very helpfull, but I'll
see if I can point at a different server that isn't so busy so incase something is getting lost in a flood of messages.

Sean Ennis
Thursday, June 24, 2004

We just hit this very problem last night.  We ended up with 50 copies of the same email sucked into the case (with all of the large attachments) until we discovered the issue, and manually deleted the message from our POP server.

So, is there a fix and/or workaround for this problem?  Also, can we somehow remove the 49 duplicate entries on the case?

Joe Acunzo
Saturday, October 30, 2004

*  Recent Topics

*  Fog Creek Home