Closed Bug 323193 Opened 19 years ago Closed 15 years ago

SMTP Server Timeout set too low during Send (from client IP with invalid reverse DNS)

Categories

(Thunderbird :: Message Compose Window, defect)

x86
Windows XP
defect
Not set
major

Tracking

(Not tracked)

RESOLVED INCOMPLETE

People

(Reporter: joerg, Unassigned)

Details

(Whiteboard: closeme 2009-09-09)

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8) Gecko/20051111 Firefox/1.5
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8) Gecko/20051111 Firefox/1.5


After upgrading to Thunderbird 1.5, sending mail via an SMTP server from a client IP with invalid reverse DNS times out before the send can compelete.  After composing the mail and clicking send, the older version fo thunderbird would hang for 20-40 seconds before eventually sending the message.  This was because of the DNS timeout looking up and invalid reverse IP.  However, in 1.5, this timeout seems to be lower, so instead of eventually sending the message, it simply say it cannot talk to the SMTP server.

I have confirmed that this pre 1.5 still works.

Reproducible: Always

Steps to Reproduce:
1. Configure you client on an IP with invalid reverse DNS (i.e.  reverse of IP results in querying non-existant domain
2. Compose and send a message
3. It will pause, then fail

Actual Results:  
Thunderbird says it cannot talk to the SMTP server, but in actuality, it does "connect" I can see that in netstat.

Expected Results:  
It should have had a higher timeout set so that after the server times out trying to verify reverse DNS, the client is still around attempting to send and succeeds.  This is how pre 1.5 worked.

Thanks for the great work.
Did you try to increase timeout value?
Go Tools/Options/Advanced/General, then increase "connection timeout" value(60sec is default).
This can be changed by modification of mailnews.tcptimeout value too.
(thru "config editor" in Tools/Options/Advanced/General panel) 
I did try that.  I upped it to 120.  The weird thing is that the initial "Sending Message" popup says "Connected to ..."  so unless that message is in error it actually is connected.

Secondly, the time between ther initial popup and the error message is still only 60 seconds.  I have restarted Thunderbird as well as windows, which leads me to believe the timeout setting may only affect receive attempts.

Telnetting to port 25 results in a connection.
Similar issue to Bug 234320 or Bug 312600?
Protocol log maybe will help problem analysis by developers.

See http://www.mozilla.org/quality/mailnews/mail-troubleshoot.html#smtp
and replace NSPR_LOG_MODULES=protocol:5 part by
  NSPR_LOG_MODULES=SMTP:5,nsSocketTransport:5,nsHostResolver:5
Then attach the log file to this bug.
Please replace sensitive information(such as mail address) in log file by random string before open the log to public.
(I think changing back of mailnews.tcptimeout to default is better for problem analysis.) 
QA Contact: message-compose
Joerg indicated he would provide a log - but don't know if he is still getting mail.
Summary: SMTP Server Timeout set too low during Send → SMTP Server Timeout set too low during Send (from client IP with invalid reverse DNS)
I've got a similar SMTP timeout problem where the first attempt to send a message  after starting Thunderbird or the first one after a longish stretch of inactivity always fails, however, setting the mail.server.server1.timeout property to a higher value (e.g. 120, 300, ...) doesn't help because Thunderbird keeps reverting it to 29.(!)  This sounds like an simple obvious bug in the property handling code that ought to be easy to track down.
(In reply to comment #5)
> setting the mail.server.server1.timeout property to a higher value (e.g. 120, 300, ...) doesn't help because Thunderbird keeps reverting it to 29.(!)

How can mail.server.serverX.timeout have relation to problem of SMTP(=this bug)?
If timeout value for specific SMTP server exists, I think it should be mail.smtpserver.smtpX.timeout or something.

I could observe phenomenon of "mail.server.serverX.timeout is reset to 29" with dummy IMAP account of dummy IMAP server, using Seamonkey 1.1.7.
 (Couldn't observe reset to 29 with dummy POP3 account.)
 (Couldn't observe reset of mail.smtpserver.smtpX.timeout for dummy SMTP.)  
 (I don't know whether mail.smtpserver.smtpX.timeout is used or not.)
But, when trunk(2007/12/19 build), I couldn't see phenomenon of the reset with dummy IMAP server.

To Tom Culliton:  
If problem can be reproduced with latest-trunk nightly, please open separate bug, after search bugzilla.mozilla.org well for already opened or DUP'ed or already FIXED bugs.
Assignee: mscott → nobody
Reporter could answer to comment #3, please?

Tom Culliton could answer to comment #6, please?
Whiteboard: closeme 2009-09-09
(In reply to comment #5)
> setting the mail.server.server1.timeout property to a higher value (e.g. 120, 300, ...)
> doesn't help because Thunderbird keeps reverting it to 29.(!)

mail.server.serverX.timeout was for interval of IDLE command. So, if larger than 29, it's set to 29, because IMAP client should issue ILDE again withn 30 seconds.
If you are complaining about "Tb keeps reverting it to 29", it's INVALID.
RESO INCO due to lack of response to last question. If you feel this change was made in error, please respond to this bug with your reasons why.
Status: UNCONFIRMED → RESOLVED
Closed: 15 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.