Why not open port 110 (or a non-standard port like 8110) to your mail server so your offsite users can get the mail directly from your server?

If you don't have a static IP or domain name you could set up a 'dynamic dns' one.

If you are doing this you could also open an SMTP port so all the mail goes out through your server as well. (would recommend a non-standard port if you are not wanting to receive mail by SMTP) 

Another option is to set up a forward or alias for these users, but it would have to point to a non-local address.

I will complain :-)

 

Well.

I have the same problem here. From the moment I first time installed Mercury, year and so.

And it is a WONDERFUL PRODUCT but this little annoyance is just ... annoying :-)

When I start Mercury it fetches my emails via POP3 from outside mail server and delivers them locally to be read by IMAP client.

Of

FIRST fetch, no matter how many emails fetches - app. 1/4 messages get

stucked by "Transient error - job deffered for later processing".

Others are delivered but this 1/4 stays in "Pending" state. After some

time(?) they get delivered to. Don't know about this "time" parameter

but I would say it is something about 20 mins or so.

At that time many files in QUEUE have  this "File rename failure for xxxxxx" message.

Now, I have NO AV program, NO any other  program that would access files and interfere with mercury. No NOTHING.

Frankly, I would plainly say there is a bug somewhere so please fix it and make this great software even better.

 

Thank you.

 

Your session log should show something like:

[quote]

FOR CRAM-MD5

13:48:28.243: >> EHLO [192.168.0.3]<cr><lf>
13:48:28.246: << 250-mail.mydomain Hello [192.168.0.3]; ESMTPs are:<cr><lf>250-TIME<cr><lf>
13:48:28.249: << 250-AUTH CRAM-MD5 LOGIN<cr><lf>
13:48:28.251: << 250-AUTH=LOGIN<cr><lf>
13:48:28.253: << 250 HELP<cr><lf>
13:48:28.265: >> AUTH CRAM-MD5<cr><lf>
13:48:28.268: << 334 -munged challenge hash-<cr><lf>
13:48:28.282: >> -munged response hash-<cr><lf>      << Base64 decode this line to get the username + morehash
13:48:28.282: << 235 Authentication successful.<cr><lf>
13:48:28.295: >> MAIL FROM:<user@mydomain><cr><lf>

 

FOR FOR AUTH=LOGIN

14:08:19.268: >> AUTH LOGIN<cr><lf>
14:08:19.275: << 334 VXNlcm5hbWU6<cr><lf>
14:08:19.369: >> -munged base64 username<cr><lf>
14:08:19.369: << 334 UGFzc3dvcmQ6<cr><lf>
14:08:20.461: >> munged base64 password<cr><lf>
14:08:20.461: << 235 Authentication successful.<cr><lf>

 [/quote]

If you get no AUTH lines in your session log, then Thomas is right, and you are a (partially) open relay  [:O]

[quote user="Judah Davis"]Can Pegasus Mail operate as an IMAP client?  We have only used it as a POP3 client.[/quote]

 

Yes, it's had IMAP4 capability for some time now.  Use Tools | IMAP profiles to setup the IMAP4 accounts. 

The outgoing e-mail still goes to Mercury, although the definition in Pegasus correctly specifies my ISP's mail server. Even disabling MercuryC doesn't address this - e-mail still goes to Mercury.

I have checked the "Use in preference to LAN mailer" box in Pegasus, and that seems to have done the trick.

Thanks for your help, Rolf.

cheers,

david

I have in the past and am currently now running two instances of Mercury/32 looking at the same mail store delivering the mail.  Generally there is no problem at all and we are processing about 500 messages a day.  Just make sure the only thing common to the two instances are the same mail store.  The must have a completely different directory structure except for the mail directories.

 Edit:  FWIW, multiple instances of Mercury/32 on the same system is not a problem.  You'll have to have separate IP addresses for each version to do SMTP delivery on port 25.

 

 

hm the checkboxes was already not ticked but i have enable the copy-to postmaster function if there are errors.

i try to check it again if i get the loop-discarded-info again

Thomas & Dilbert,

Thanks for taking the time to correct my naivety in this matter. I'll take a look at the POP3 connectors, which is where Thomas directed me to in the first place.

Regards,
Richard

> Hi again
>
> Yes, 110.66.159.212.is the IP of our mail server and
> mail.mischiefkids.co.uk is our own domain name.
> I did set the controls to disallow everything as you said Thomas,
> and above the disallow I put the mailfoundry range of addresses.
> As some valid mail could not then get through (e.g. coming from
> Yahoo) I had to remove the disallow.

Ok, now that we have the info I think we can see the problem.  Your domain is really
mischiefkids.co.uk and this mail is being routed via the mailfoundry address.  The name
mail.mischiefkids.co.uk is the name assigned to your mail server.  If you sent mail from Yahoo
to user@mischiefkids.co.uk it should have gone through properly.  People should not be
sending to user@mail.mischiefkids.co.uk unless they are spammers and you want them
blocked.  Mail from Yahoo to your domain addressed to user@mischiefkids.co.uk should be
sent via the mailfoundry hosts.  

Answer Section:
    mischiefkids.co.uk, A, 89.145.78.228
    mischiefkids.co.uk, MX, 10, mx1.mailfoundry.com
    mischiefkids.co.uk, MX, 10, mx2.mailfoundry.com

Answer Section:
    mx1.mailfoundry.com, A, 66.18.18.112

Answer Section:
    mx2.mailfoundry.com, A, 66.18.18.113

The domain mail.mischiefkids.co.uk though has no MX records at all.  Any mail to
user@mail.mischiefkids.co.uk is going direct to the server.  

Answer Section:
    mail.mischiefkids.co.uk, A, 212.159.66.110

You are trying to block people from sending mail direct to the server and this address should
not be accepting mail from anything but the mailfoundry MX hosts.


>
> It would seem that Dilbert is correct (thanks for that Dilbert) and
> I will just have to be patient and keep trying the disallow settings
> every few weeks.

Not really, you need to get a tool to check the DNS records for your domain every once in a
while to verify what is being propagated to the world. The site CheckDNS is a good place to
start.  FWIW, I use the free tool Cyberkit to do this sort of thing.

> Thank you for your input and confirming that my DNS settings are
> correct.
> Lynn

Do a 'file compare' on your two mercury folders and play 'spot the difference' [:D]

 

I am using Mercury Content Control and Filtering and have been very pleased to reduce my Spam-count from 200+ to 0/1/2 a day, with no false positives.  However, although it isn't critical, it would be nice to be able to get rid of the last few.  One common feature of most of the remaining spam is that a http URL appears by itself on a separate line in the message body.  Is there any way that Mercury can recognize this?  Even being able to test that a line begins with "http" would help.

 Thank you

GordonM

[Edit} It seems that I have now solved my own problem. 

 if BODY MATCHES "/bhttp://*.com/B" weight 51 tag "http"

seems to work for decting http URLs on isloated lines (at least for .com addresses). 

The local topology and routing is essential, as where you place dns and how you point dns records to avoid using the uplink.

You should start a new thread about this, and be a bit more clear about the network topology first.

Yes, I thing a dynamic IP address with a rDNS liker that is going to cause you problems.  Try using MercuryC with your ISPs SMTP host as a relay host and see what happens.

Thanks, that's good to hear.

After my previous post I reread the manual an discovered synonyms. I think that using synonyms in the form of

<internet email address with periods> == <local username without periods>

is an acceptable workaround.

I find POPFile and it's magnets a bit easier to use but if you are already running at 98% with Spamhalter I really do not think it's will help you all that much.

 

Dave - Yes, I think the rules are working correctly for me.  

I am using:

if HEADER

"Received" CONTAINS ".ru ([" weight 51 tag "foreign"

... so no asterisk.  I don't think that * should be used in simple string tests, only in regular expressions (unless one is trying to match a * in a string).

Other rules that I am finding useful are of the form: 

if HEADER "Received" CONTAINS ".pt[" weight 51 tag "foreign"

if HEADER

"From" CONTAINS ".lv>" weight 51 tag "foreign"

if SENDER CONTAINS

"*.ro>" weight 51 tag "foreign"

if HEADER

"Received" CONTAINS ".do (" weight 51 tag "foreign"

if HEADER

"Received" CONTAINS ".kh)" weight 51 tag "foreign"

Gordon

 

Thanks.

I'll check my logs and make sure it's working as expected.

It's just a zip file with mxredir.txt & mxredir.dll

EDIT: Oops, that was my zip file, it's just the .dll available here (http://www.xs4all.nl/~fenke/files/mxredir.dll). The text file was a post from the mailing list. (copied below)

 

It seems fairly basic / beta but I tested it on 4.01b and it worked as advertised. Have not tried it on 4.5x but it should work.

Note the cautions about dropping undeliverable mail, I would suggest an 'Archive' filtering rule for that domain mail for backup purposes in case the network (or more likely Exchange [:P]) goes down.

 

 BEGIN MXREDIR.TXT

Have you thought of using a MX type operation here using a daemon?
Mercury/32  does not yet have the capability of forwarding all mail for a domain
off to another server.  That said, Frans Meijer has written a small daemon that
can do this.  I've tested this and it seems to work just fine.  You might want to
setup a Mercury/32 "Always" filter to send all of the mail for this domain off to
one of your local mail accounts so you'll not lose mail.

---------------------------------- quote ------------------------------------------------
A first version - for testing, not production use - is at

http://www.xs4all.nl/~fenke/files/


The actual daemon to use is mxredir.dll the rest are the sources to build it, for
which you also need daemon.h from the Mercury distribution and Borlands free
C++ compiler (should not be hard to port to other compilers though - mind the
alignment issues). 

Put the dll in a convenient directory, for instance the Mercury directory, assume
C:\Mercury for the remaining text and 192.168.2.3 is the ip address of the
receiving MTA. 

In Mercury Core config, local domains tab add a new domain. In the Edit
domain entry dialog fill in the local server and internet domain:

Local host or server: daemon:C:\Mercury\mxredir.dll;[192.168.2.3]
Internet name: test.local

(You can off course another domain name and ip address)

Messages send to <any-mailbox@test.local> should now be processed by the
daemon, you can see this in the core-process window (Job ... from ... To:
daemon:C:\M ...etc...) 


The daemon writes to a log file 'redirector.log' in the Mercury
directory.


Caution.

Mercury expects the domain-daemon to handle the message, it hands the job
over, then discards the original. If, for whatever reason, the daemon can not
forward the message, it is lost. A backup mechanism should be added, for
instance writing the message to a file somewhere, or dropping in a user mailbox
(though if it can't create the forward message that might be a problem to). Your
input is welcome. 

I feel there should be some more validations, for instance on the parameter
passed, maybe on the rcpt addresses. Perhaps not use a logfile (another point of
failure) and certainly not abort if it can't be opened. More? 

The current daemon will likely not work if the _source_ is a domain-popbox, I
don't know (yet) how the rcpt to address will look in the job-info passed to the
daemon, the daemon expects <mailbox@domain> 

As said, this is not (yet) for production. If anyone gives it a test-run, let me know
how it works, problems and change requests. (post here or to reply-to or if that
fails: <frans.meijer@xs4all.nl>) 

------------------------ end quote ---------------------------------------------------

END MXREDIR.TXT 

[quote user="blake"]

Hello,

    I'm  using Mercury/32 on Ubuntu under WINE, and I'm pretty sure I've screwed things up. [:(]

   Right now, the problem is that I'm getting this error from my client (Opera):

Message does not exist! [Server response:-ERR file error.]


    I did delete a bunch of files... How do I track this down?

    ===Blake===
 

[/quote]

 

That "I did delete" makes me worry.  Might be that you've deleted a necessary file or even directory.   Not sure though how you are accessing you files and when do yuo get the error message.  If via POP3 then I'd turn on session logging in MercuryP to see what's going on between the server and the clist.

 

Thanks.

That worked !!