Forums & Blog

Hi Guys,

 I feel this is a bug that needs some attention,

I see emails sitting in my spool queues waiting for delivery, they have been in the queues (pre SM4.1) for 4 hrs and retrys sitting at 4 and (Post SM4.1 I was hoping it would have been resolved by this release) 8 retrys.  If I do external tests to these email addresses which are retrying, some are for domains that dont exist, others are for the recipients email account doesnt exist, or their accounts are full.

for example this email was received, it is in the smtp log as being received, but there is nothing in the delivery log mentioning it was even tried to be delivered. Here is the SMTP log entry for this message.

14:03:07 [xxxxxxxx][2344757] rsp: 220 mail.xxxxx.co.za 
14:03:07 [xxxxxxxx][2344757] connected at 3/20/2007 2:03:07 PM
14:03:07 [xxxxxxxx][2344757] cmd: EHLO xxxxx.xxxxx.co.za
14:03:07 [xxxxxxxx][2344757] rsp: 250-webxxxxx Hello [xxxxxxxx] 250-SIZE 20971520 250-AUTH LOGIN CRAM-MD5 250 OK
14:03:07 [xxxxxxxx][2344757] cmd: AUTH LOGIN bWFpbEBob3N0d29yeC5jby56YQ==
14:03:07 [xxxxxxxx][2344757] rsp: 334 UGFzc3dvcmQ6
14:03:07 [xxxxxxxx][2344757] rsp: 235 Authentication successful
14:03:07 [xxxxxxxx][2344757] Authenticated as mail@xxxxxxx.co.za
14:03:07 [xxxxxxxx][2344757] cmd: MAIL FROM:<notify@xxxxxxx.co.za> SIZE=1323
14:03:07 [xxxxxxxx][2344757] rsp: 250 OK <notify@xxxxxxxxxx.co.za> Sender ok
14:03:07 [xxxxxxxx][2344757] cmd: RCPT TO:<lm@easypayroll.co.za>
14:03:07 [xxxxxxxx][2344757] rsp: 250 OK <lm@easypayroll.co.za> Recipient ok
14:03:07 [xxxxxxxx][2344757] cmd: DATA
14:03:07 [xxxxxxxx][2344757] rsp: 354 Start mail input; end with <CRLF>.<CRLF>
14:03:07 [xxxxxxxx][2344757] rsp: 250 OK
14:03:07 [xxxxxxxx][2344757] Data transfer succeeded, writing mail to 118883533072.eml
14:03:07 [xxxxxxxx][2344757] cmd: QUIT
14:03:07 [xxxxxxxx][2344757] rsp: 221 Service closing transmission channel
14:03:07 [xxxxxxxx][2344757] disconnected at 3/20/2007 2:03:07 PM 

 And here is the header of the email sitting in the queues retry no 6 (mine are set to 15,15,30,60) There is no mention of the server even trying to deliver this message in the delivery log, if I do a nslookup on the mail server it says the domain does not exist. However this still sits in the queue until the timeouts expire before replying to the user that the meassage was undeliverable because the domain does not exist.

Return-Path: <notify@xxxxx.co.za>
Received: from webxxx.xxxxxx.co.za [xxxxxxxxx] by mail.xxxxxxx.co.za with SMTP;
Tue, 20 Mar 2007 14:03:07 +0200
Date: Tue, 20 Mar 2007 14:02:12 +0200
Subject: Hello
To: lm@easypayroll.co.za
MIME-Version: 1.0
Content-type: text/html; charset=iso-8859-1
From: dreammatch.co.za <notify@xxxxx.co.za>
Return-Path: notify@xxxxx.co.za
Message-ID: <WEBxxxxxxoygXLsxpTl00000022@Webxxxxx.xxxxx.co.za>
X-OriginalArrivalTime: 20 Mar 2007 12:02:12.0484 (UTC) FILETIME=[9814D440:01C76AE7]
 

only once the full amount of retries fail does it come back with this error

Could not deliver message to the following recipient(s):

Failed Recipient: lm@easypayroll.co.za
Reason: Failed four attempts to connect to the recipients mail server. No DNS information was found for the 'easypayroll.co.za' domain.

Is this by design or an error, I want to setup my number of retries to cover a 24 hour period, but if a client sends an email to a domain that doesnt exist exixst it means the user must wait 24hr or longer to get a reply telling them that the domain is invalid... Surely if dns for the domain does not exist this should be bounced immediately. if the domain exists and dns is faulty for a remote server this surely should also bounce immediately and not have to wait 24hrs.

Another one, telnet to this users server and sending to her from the sender works fine, smtp diag works fine as can be seen below, however SM cannot deliver it to the user, as can be seen by the respionse email.

This is an automatically generated Delivery Status Notification.

Delivery to the following recipients failed.

nikkif@finlay.co.za 

Non-authoritative answer:

Server:  ns02.xxxxxxxxx.co.za
Address:  xx.xx.xx.xx

finlay.co.za MX preference = 20, mail exchanger = smtp.finlay.co.za
finlay.co.za MX preference = 10, mail exchanger = dynamic.finlay.co.za

smtp.finlay.co.za internet address = 196.25.28.154
dynamic.finlay.co.za internet address = 41.241.157.190

Searching for Exchange external DNS settings.
Computer name is Popeye.
Failed to connect to the domain controller. Error: 80004005

Checking SOA for finlay.co.za.
Checking external DNS servers.
Checking internal DNS servers.
SOA serial number match: Passed.

Checking local domain records.
Checking MX records using TCP: xxxxxx.co.za.
Checking MX records using UDP: xxxxxx.co.za.
Both TCP and UDP queries succeeded. Local DNS test passed.

Checking remote domain records.
Checking MX records using TCP: finlay.co.za.
Checking MX records using UDP: finlay.co.za.
Both TCP and UDP queries succeeded. Remote DNS test passed.

Checking MX servers listed for nikkif@finlay.co.za.
Connecting to dynamic.finlay.co.za [41.241.157.190] on port 25.
Successfully connected to dynamic.finlay.co.za.
Connecting to smtp.finlay.co.za [196.25.28.154] on port 25.
Successfully connected to smtp.finlay.co.za.
 

Nothing in the Delivery log for this email either...

 

The retry when DNS info can not be found is intended. DNS may not be available for an amount of time, and it shouldn't give up if that is the case.

 

What was the bounce message for the second one?

Ok but my question is,

If dns is down, and an email cant be delivered because of this situation, surly rather return the message as undeliverable to notify the user that there is a problem rather..

If the email is important and not delived due to a dns problem, the chances are that the end user is unaware of the dns issues and will only find out after a long delay that there is an issue, in the meanwhile the user is assuming that the email has been successfully delivered only to find out that they typed in the wrong domain name by mistake..... its an oopsie that im seeing far too regulary at the moment and users are bitching that why didnt the server return the email sooner, but now that im on SM4.1 I intend on moving my retries up a lot longer... So its just going to make this issue a whole lot bigger. Its a catch 22 I know cant please everyone all the time, maybe cant the guys put in an option tochoose if the mail server should retry the emails or return it.

The second email.. well that is just it I cannot find an entry in the delivery log for that email. no failure whatsoever, however the email bounced as you can see..

 

 

Ok Baaack,

 Definately something funny going on here.. thses also stay in the queues and are definately a problem the user does not exist but the message stay in the queues ubntil it times out eventually after all the retries and says user does not exist.

20:30:08 [86352] Sending remote mail for antoinette@xxxxxx.co.za
20:30:08 [86352] Connecting to 196.25.211.166
20:30:08 [86352] Connection to 196.25.211.166 succeeded
20:30:08 [86352] RSP: 220 webmail.co.za ESMTP Welcome to the Webmail.co.za SMTP Server.
20:30:08 [86352] CMD: EHLO mail.xxxxxx.co.za
20:30:08 [86352] RSP: 250-webmail.co.za Welcome to the Webmail.co.za SMTP Server.
20:30:08 [86352] RSP: 250-PIPELINING
20:30:08 [86352] RSP: 250-SIZE 11048576
20:30:08 [86352] RSP: 250 8BITMIME
20:30:08 [86352] CMD: MAIL FROM:<antoinette@xxxxxxx.co.za> SIZE=1086887
20:30:12 [86352] RSP: 250 ok
20:30:12 [86352] CMD: RCPT TO:<xxxxxx@webmail.co.za>
20:30:12 [86352] RSP: 554 Sorry, no mailbox here by that name. (#5.1.1)
20:30:12 [86352] CMD: QUIT
20:30:17 [86352] The smtp session has timed out.
20:30:17 [86352] Connecting to 196.25.211.170
20:30:19 [86352] Binding to local ip 'xx.xx.xx.xx' failed.  Switched binding to primary local ip.
20:30:19 [86352] Connection to 196.25.211.170 failed

 

 

smtpdiag.exe result 

Received:
554 Sorry, no mailbox here by that name. (#5.1.1)

Error: Expected "250". Server rejected the recipient address.
Failed to submit mail to mail.telkomsa.net.

Log for entry

 13:22:08 [86235] Sending remote mail for notify@xxxxxxxx.co.za
13:22:08 [86235] Connecting to 196.25.211.150
13:22:08 [86235] Connection to 196.25.211.150 succeeded
13:22:08 [86235] RSP: 220 telkomsa.net ESMTP Welcome to the TelkomSA SMTP Server.
13:22:08 [86235] CMD: EHLO mail.hostworx.co.za
13:22:08 [86235] RSP: 250-telkomsa.net Welcome to the TelkomSA SMTP Server.
13:22:08 [86235] RSP: 250-PIPELINING
13:22:08 [86235] RSP: 250-SIZE 11048576
13:22:08 [86235] RSP: 250-DATAZ
13:22:08 [86235] RSP: 250 8BITMIME
13:22:08 [86235] CMD: MAIL FROM:<notify@xxxxxx.co.za> SIZE=1690
13:22:32 [86235] RSP: 250 ok
13:22:32 [86235] CMD: RCPT TO:<xxxxxxxx@telkomsa.net>
13:22:32 [86235] RSP: 554 Sorry, no mailbox here by that name. (#5.1.1)
13:22:32 [86235] CMD: QUIT
13:22:37 [86235] The smtp session has timed out.

13:37:37 [86235] Sending remote mail for notify@xxxxxxx.co.za
13:37:37 [86235] Connecting to 196.25.211.150
13:37:37 [86235] Connection to 196.25.211.150 succeeded
13:37:37 [86235] RSP: 220 telkomsa.net ESMTP Welcome to the TelkomSA SMTP Server.
13:37:37 [86235] CMD: EHLO mail.hostworx.co.za
13:37:37 [86235] RSP: 250-telkomsa.net Welcome to the TelkomSA SMTP Server.
13:37:37 [86235] RSP: 250-PIPELINING
13:37:37 [86235] RSP: 250-SIZE 11048576
13:37:37 [86235] RSP: 250-DATAZ
13:37:37 [86235] RSP: 250 8BITMIME
13:37:37 [86235] CMD: MAIL FROM:<notify@xxxxxxx.co.za> SIZE=1690
13:38:05 [86235] RSP: 250 ok
13:38:05 [86235] CMD: RCPT TO:<xxxxxxx@telkomsa.net>
13:38:05 [86235] RSP: 554 Sorry, no mailbox here by that name. (#5.1.1)
13:38:05 [86235] CMD: QUIT
13:38:10 [86235] The smtp session has timed out.

And so it carries on so even when it receives a no such user response it ignores it and then the session times out eventually and the server then keeps on retrying it 

 

BUMP .....

 

The cause for the retries in your last post are probably because the session timed out. Even though, it got a 5xx error (which means permanent failure, and should stop the delivery), that is overshadowed by the fact that the SMTP session did not successfully complete. It looks like SmarterMail is not getting a response to the QUIT command, so it just sits there until it times out. It seems from all the log snippets you posted that the issue is common to the same mail server. If that is the case, there may be some sort of issue hampering communication between the two servers.

I just Highlighted a few of these failurtes and basically grabbed one session from a single server which retries all the time, each session is the retry from the previous failure, ie 15, 15,15,15,30,30 settings.. there are a few servers that have this issue, we have checked between ISPs whether there is a problem, ie firewall, etc but the server is ok, reverse dns, is there, pointers, etc all ok, not in span block lists, etc

 

 

Will there be a fix for this soon? It's very annoying.
We would like to have error message like this delivered right away, not after several hours.
And also for "No such user"

 -------------------------------------------------------------------------------------------------------------------------------

Could not deliver message to the following recipient(s):

Failed Recipient: xxxx@domain.com
Reason: Failed four attempts to connect to the recipients mail server. No DNS information was found for the 'domain.com' domain.

-- The header and top 20 lines of the message follows --

BUMP.. I still am having these issues, I redirected some of my domains back via my old merak server and this problem is non existant under merak, but My whole move away from merak was to move to SM4, and I dont want to rely on merak to fix stuff which should be working in SM. If a user does not exist a user does not exist, it does not need to sit in the queue for 24hrs before being returned, and neither if the dns is not available, If a server is defined in DNS and is offline, sure keep the message in the queues, but if the server cannot find a domain name in DNS return the mail to the user and tell him the domain dont exist or cannot be looked up. Why delay the inevitable by 24hrs or whatever the retry limits are.

 Some users mistype the tld ie co.az instead of co.za or ca.za instead of co.za and the mails are urgent, but they sit in the queue for 24hrs and then are returned to the user as "domain dont exist" anyway and then the users are p'd off with us because it was urgent and why did it take so long to get returned etc... In merak if the user dont exist, domain dont exist, etc it returns the mail immediately back as undeliverable. Same goes for emails that are oversize, return the email back telling the user not wait for 24hrs or whatever the interval is set for, I have my servers to retry for 24 hrs, I want to extend it to 48, but I cant because of this problem, at least give the mail admin the option to enable/disable the return of certain error codes immediately if the so choose.

Also allow the admin to have an option to remove the attachments on a returned mail.. just send back the headers and subject and error message not the whole mail.

Just My two cents worth

Waiting for a respons from ST. This is getting HOT. More and more customers are complaining.

We want "no DNS" , "no USER", "no DOMAIN" and similar, returned right away after first try.
The clients are getting very upset, when important mails are not delivered, and there are no messages to tell.
A fast patch for 4.1 would be appreciated.

Lars 

I haven't actually had complaints on this yet, but I have seen from watching the spool that some INTERNAL email addresses sit in the spool for a long time when they are misspelled also.  For example, one of my users with an email of user@domain.com emailed another user from the same domain and typed it as user2@dommmmmain.com.  I saw it in the spool for many hours before I imagine it was returned.  It seems that if it is within the internal email list it would be a LOT faster.

Hi Guys,

 Yes this is really important, I am also having customers want to leave us because if this issue. PLEASE can you make a patch to make this optional. I have got my retries setup for 24 period and these emails are only being bounced after that time... Another option is if you restart the service the emails in the queues are restarted as if they had just been delivered into the queue, so if in any urgent situation where you need to restart the service the mails are further delayed by the time they are originally in the queue already and then a further 24hrs..

 

PLEEEEEEEEEEEEEEEEEEEASe look at this urgently.. and the logging options..

 

We are getting complaints as well.

OK anothe funny to do with this delivery issue, I have clients who send out many emails to 5 or more users. Today my mail started backing up to 400 emails in my queues, apon final investigation i have found out that if the user sends the email to multiple users ie lets say 5 and the first users email address is spelt incorrect the email sits in the queue as retry, the balance of the 4 emails doesnt even get tried.

I found this out by luck when I started going through the emails from the one user, if the user resends the email and lets say the duff spelt email is the second or third or forth email adrress any email addresses before it are sent out successfully but the remail adrreesses sit in the queue.

ie  joe@namby, user1@dom.com, user2@dom2.com; user3@dom3.com  if the first address ie joe@namby no domain tld added on purpose this email with all the correct addresses will timeout and be rejected.. if joe@namby was the last address then all the addresses before it will be sent out ok..

This comes back once again to the fact of the returning of non existant addresses immediately.. not servers but user mail accounts....

eric@xxx.co.za:	Completed
salomed@xxxxdrex:	Pending
enviro@xxxx.co.za:	Pending
enquireis@xxxxx.co.za:	Pending
ericav@xxxx.co.za:	Pending

this is where the wrong address is first

salomed@xxxxdrex:	Pending
eng@xxxxxx.co.za:	Pending
engelavw@xxx.co.za:	Pending
emile@xxxxxx.co.za:	Pending
elton@xxxxx.co.za:	Pending

 

these are now sitting on 7 retries and been in the queues for 6hours

 

Bump !!!

Guys please this is an issue...