Thanks for the quick response. I have turned on session logging in the mercurys and mercurye modules. I'm not sure how to identify a problematic message in the queue.

The server is operating normally now. I'll have to wait and see if the problem reccurs. No antivirus on the server, by the way. I haven't run a disk check but there are no disk errors in the event log.

Heh, just decided to read the specs - and yeah, you're right. It seems the people who designed the specs are more concerned about the weird ways people may implement the underlying system rather than the IMAP protocol itself.

If somebody doesn't know how to implement a translation layer between a protocol and whatever storage system they're using, I seriously doubt they should be writing a mail server. I think the people who wrote the specs should've focused more on clarity rather than worrying about possible implementation quirks.
 

You are right. I'd inadvertently changed to no formatting. Don't know how. Thanks for your help.

 Regards

 
John D.

Or create a queue file and place it in the queue.

Apparently the issue isn't as solved as I hoped.

 We got several more failures over the weekend.  I was able to have one of the problem messages sent to my personal gmail account.  The attachment was in pdf format, and was 293K.  When I tried to forward that from gmail to the originally intended office address in Pegasus  I could see the message hit the SMPT server on Mercury, but it never went to the core - or to intended mailbox. 

 I spoke to the ISP (who say it is not a problem on their end) and our consultant who suggested it might be a problem with timing out.  The TCP/IP is set to 60 - should it be higher?

Also - I tried to update to v 4.01c.  I did manage to get to 4.01b (which I needed before attempting to patch to c) but I can not get the 4.01c patch to apply.  I exited Mercury.  Copied the dll file into the Mercury directory - restarted...and still show v 4.01b as the current version.  Any ideas on that?

thanks in advance

Michele

> Hi everyone,
>
>  
>
> I've recently installed Mercury Mail on a Windows XP SP2 machine. We
> are using Novell NetWare Client 4.91 for our Network.
>
> Mercury is configured correctly to work in NDS-mode and it does work
> properly. I can get my messages either with Pegasus or by using
> POP3/SMTP/IMAP/etc.
>
> Now I have set up Mercury to run as a Service with SRVANY.EXE and
> INSTSRV.EXE both from Microsoft. This still works as intended: all
> Processes are running and responding. But when I try to log in with
> telnet to test (e.g. telnet foo.example.com POP3), the process does
> answer, but when I try to login with USER usr.CONTEXT and PASS pswd, I
> get a LoginFailure. Certainly this procedure works when I do not run
> Mercury as a Service and only when I've manually logged in as a
> NetAdmin (user with sufficient rights) in Novell.
>
> In Mercury.ini i have set:
>
> [NDS]
> UserID: Netadminfoo.RWTH
> Password: barpw
>
> Is this right?
>
> Why does Mercury not login itself with the give User and is it
> possible to run Mercury as a service and with NetWare support?

Hard to say why it's not authenticating to the Netware host but IIRC the service is running as system and does not authenticate to the Netware host.  I use NT Wrapper to connect to my bindery host since it has the capability of running as a specific user to make the Netware connection.

The NT Wrapper allows standard Win32 applications or scripts to be run as a Windows NT/2000/XP/2003 Service.  

Features:
    ·    Easy configuration thru a GUI and simple INI files.  
    ·    Prioritization of sub-processes.  
    ·    Custom environments.  
    ·    CPU binding  
    ·    Redirecting of Stdout/Stderr to file  
    ·    Logging to the event log and to disk.  
    ·    The capability to run multiple applications in a
          single NT Wrapper service instance.  
    ·    Monitoring of a service in the sys-tray.  

http://www.duodata.de/ntwrapper/



>
>  
>
> Thanks for help
> Sincerely
> Horat
>  

Hello,

Mercury is used on this Win2003 Server hostname DME to send out mail from a Unix host.

DME is part of a .local domain.

The outgoing email from the Unix host always uses the same From address.

Some  jobs are sent OK but some show message loop detected - message discarded.

What should be the correct entry in the  edit local domain page for local host and internet name?

I read the help file but wasn't sure because of the .local hostname. 

Thanks,

Oded 

Hello!

What is the subject to trigger that content control rule?

I think what you want to find is a line like
Subject: [SPAM] Buy this!
i.e., a "Subject:"-line your ISP has added the "[SPAM]"-tag to.

If this is what your content control is to find, then I suggest the following content control rule:
IF SUBJECT MATCHES "[[]spam]*" WEIGHT 51 TAG "Flagged by NTL"
Note that the leading bracket in "[SPAM]" has to be masked. When used in a regular expression, "[" is considered a metacharacter; if you want to use it as a normal character, "[" has to be masked, e.g. placed within "[ ... ]".

Perhaps, you also need a leading blank space before the Spam-tag. Then, the rule could look like this:
IF SUBJECT MATCHES "[[]spam]*"
OR SUBJECT MATCHES " +[[]spam]*"
WEIGHT 51
TAG "Flagged by NTL"

[quote user="nijinski"] 116.25.87.49 this is my ip address according to whatip.org[/quote]

Rolf wrote: "I think you first need to verify that  220.112.193.167 is your current IP"

I can see your IP changes frequently. Currently your MX points to conneally.net, and conneally.net has an A record for 220.112.193.167. This would work, as long as that was your current IP, which it isn't. So somehow your dyndns isn't updating your current A record for conneally.net

Regarding the port forwarding:

First test and verify that you can reach MercuryS from your own network, by using telnet as you did: . If you do not get a response here, something within your machine is blocking inbound traffic (f.ex. windows firewall or another port blocking software)

Secondly, test and verify that you can telnet from the outside through your ADSL modem port forward based on IP number, telnet:116.25.87.49:25

Lastly, check that the DNS matches your current IP. You can easily do this with nslookup from a command  window. And when this works, telnet:conneally.net:25 and you should be through the hardest parts.

[quote user="Methuselah"]

Hi there.

I was trying to implement an autoreply using template.
( I chose failure.mer  for starters)

I tested by matching to a known webmail address, twice.

I think it worked, because I saw the Mercury Core Process display showing a message being created by postmaster@myserver.com going back to the address I sent from.
There are also 2 jobs listed under Ready: in the Outgoing Jobs: line.

  
Thing is, they never leave - no response to the sender over an hour later.

1.  Where can I view these jobs ?

2.  How can I get them to go - I guess there is some config thing I have missed.

Robin.

 [/quote]

 I assume you are replying to a non-local address and you saw core create the message for the external email address.  You should see these messages in the outbound queue as a MO*.QCF/QDF file pair.  This is assuming that MercuryC/E has not processed the outbound mail.  What do the system and the MercuryC/E logs show?

Hi Chris, we have the same setup as you.

Mailbee WebMail is set to default for all new users to use IMAP.

The outlook clients and PDA clients are beyond our control, but to the outside world we do not expose IMAP - so they are bound to use POP3.

This setup works perfect in 95% of all cases where the users are either in the office or on the road.

For the reminder 5% we tell them to set the POP3 setting to leave the messages on the server for x number of days. Some users share computers, some exchange and answer inbound general mail from different computers and for most clients this setting does not duplicate messages.

Regarding Mailbee, the setup is also to make duplicate copies within the database. You can set up Mailbee to synchronize in direct mode, full message, headers or not - meaning copying inbound messages and treat them only within the webmail client. Currently we're working with Afterlogic in the beta process of v4.1 in ASP.NET.

There's a nifty little tool named poppass 1.05 for win32 written by Michael John Phillips / Vortex Software which I'm using since years. Unfortunately it's page is down and I couldn't find it anywhere else on the net so I'm providing it at my server: 

Best regards

Nico

> Hello.
>
> We'd like to move from Sendmail(unix) to  Mercury(win2003).
>
>  We do not need to migrate existing mail, because it is a training lab.
>
> We'd like to create and delete  Mailboxes via Script. How can that be done?
>
> Pointers to existing manuals very welcome
>
> Bye
>

Should be pretty straight forward to modify the pmail.usr file and make the subdirectories for the users.  The default pmail.ini and passwd.pm files are also all self explanatory. The only thing that you probably need to do is use the reload users command if you are trying to do this with Mercury/32 running.

*   The mail server now has its own password file, created and
    managed on the "Mail server" configuration page. This is not
    covered in the help yet, but you can create any number of mail
    server passwords that can be used in a "PASSWORD" command to
    handle the new operations described below. If you have no
    password file, then no password will ever succeed and the new
    operations will fail automatically - hence the operation is secure by
    default. Mail server passwords are simply one-per-line (like
    everywhere else) and cannot contain the character ';' (semi-colon).

*   A new mail server command, "KILLFILE" has been added. This
    command allows you to manipulate the MercuryS killfile by mail. At
    present, the only implemented command is "KILLFILE ADD", which
    adds the parameter you supply to the MercuryS killfile. An example
    is at the end of this message.

*   A new mail server command, "RELOAD" has been added. This
    command allows you to force Mercury to re-read certain
    configuration files by mail. At present, the only implemented
    command is "RELOAD USERS", which takes no parameters, and
    tells Mercury to reload the user database, PMAIL.USR.

Both of these commands need to be preceded by a PASSWORD
command specifying a valid password from the password file created in
the mail server configuration dialog.

Examples (these examples assume that the password 'foo4u' is
specified in the Mail Server password file):

* To add *@163.com to the MercuryS killfile, issue these commands:

      PASSWORD foo4u
      KILLFILE ADD *@163.com

* To force Mercury to reload its user database, issue these commands:

      PASSWORD foo4u
      RELOAD USERS

The commands are not case sensitive - I've shown them in uppercase only to emphasize them.

I plan to add more commands of this kind - administrative in nature - and will be adding a level of discrimination to the password file to allow passwords to work only for certain types of operation (hence the reason why ';' cannot be used in a password).

i think it is best for me to review my situation in view of the long post i have created. with this in mind i would like to begin again. watch for another NEWBIE post.

 thanks in advance

gjc 

Users are setup using Configuration | Manage local users.  The postmaster is specified as one of the local users so error messages have a place to go.  The domains in the server are setup using Configuration | Mercury core | Local domains.  The domain of the server is set in Configuration | Mercury core "Internet name for this system"  You might want to checkout http://kbase.pmail.gen.nz/viewfull.cfm?ObjectID=C1523317-4EC6-11D4-8B6B0008C709E5EC to see if some of your basic questions could be answered via the knowledgebase.

On 22 Nov 2007 Pegasus Mail & Mercury - Automated Email <NoReply@praktit.se> wrote:

> Hi guys.
>
>  A little bit of both.
>
> The Shared Mailbox thing:
> I setup mail to all users ( i.e *@agrilibrium.co.za) to be forwarded
> to agrilibrium@iburst.co.za.
> (This is done from the domain host cPanel).  The split is done after
> retrieval by Mercury.
> My reasoning was to avoid defining a mailbox for each of the users
> at the domain level.
> I simply administer locally via Mercury.
>
> Mail from this person (of the Exim server) is processed correctly
> when they are on TO: or CC:
> It's only when on BCC that it doesn't work as the To: header shows
> <Undisclosed Recipients>.
>
> I have no idea whether I am processing Envelope_to: before any To:

The X-Envelope-To: is being added by Mercury/32 when delivering the message to a new mail directory.  If your forwarder has not added the original RCPT TO: addresses the the mail when it's being rewritten and forwarded then you'll never be able to determine an original Bcc: address since it's only in the SMTP addresses in most cases.  The forwarder normally should add this as a Deliver-To: or similar header that you can use in delivery.
   
> Where would I check this ?

[quote user="dpulver"]

Hello,

 i want to use the Mercury Server as a IMAP Server. I readed, that there is a difference for performance between the maildir and the mbox

format. Which format supports the mercury server ?

Is it possible to manage many thousand mails per user as IMAP with a good performance?

Thanks

[/quote]

Neither, the mailbox format is the Pegasus Mail file and folder format.  I use Mercury/32 IMAP4 with SquirrelMail to provide web mail services and find it works quite well over a broadband connection.  Thousands of message may or may not be a problem.  I open folders with a couple thousand messages with no problem; I generally shy away from opening folders with over 10K messages because it's slow.  My personal account has a few hundred separate folders and many thousands of messages.  Whether it will work for you or not is really based on the number of concurrent users, the capability of the hardware you are using and the bandwidth.  The mail clients also come into play here, some are a lot better than others.

 FWIW, the IMAP4 protocol is a very busy protocol and does put a lot of stress on a system.

> Hello,
>
>  we are running Mercury32 4.51 on Windows XP and Windows 2000
> Server as IMAP-Servers and Mercury Pop3 Server module 1.48 on
> Netware 6.5
>
> On both IMAP- and Pop3-Servers the Client has to login with
> the full Novell-Name like .User.Context.Tree to authenticate
> with Novell NDS.
>
> To make things easier for our Users is there a way to make a
> login with only the Username (without the Container and Tree)
> possible?

This is what I see in the readme.nds for POP3, it's probably good for IMAP4 as well:

11: POP3 server considerations

Under NDS, it is possible to have exactly the same username defined in
multiple contexts in your NDS tree. This can pose a problem for MERCNDSP,
the POP3 server, in that when presented with a name like "david", it may
be unable to tell which of several possible "david"s in your NDS tree is
the proper user. To get around this, you can take one of two approaches:

  * Require your users to give a full NDS distinguished name when
    logging in to the server (for example, "david.dev.pmail").

  * Create a POP3 alias file, POPALIAS.MER, in SYS:SYSTEM. This file
    simply equates a simple username (like "david") to the full NDS
    format used for logging in (like "david.dev.pmail"). POPALIAS.MER
    is a simple text file where each line has the form:

        <pop3_login_name> = <nds_username>

    As an example, using the names above, you would have the line

        david = david.dev.pmail

    NDS usernames must be expressed as full paths relative to the root
    of your NDS tree. Each line must begin hard against the left margin.
    Lines beginning with #, * or ; are regarded as comments and are
    ignored.

>
> Usernames are unique.
>
> Thanks for your help
>
>  Anke
>

 

 Thomas wrote:  "There is nothing that MercuryS, MercuryI and Mercury P can do to identity anything in the customers equipment or mail client."

 The Directive says "identify users' communication equipment or what purports to be their equipment"

 I don't think, from that wording, that there's a requirement to identify the type of equipment, or anything in it. All it requires is to identify the equipment, ie, its FQDN or IP address. I assume the second part " or what purports to be their equipment" is avoid providers being in default when they receive mail from spoofed addresses.

I agree with the above comments about blocking the rest of the world being not the smartest method to combat spam. However, if you want to use this approach for some reason you could do it by using the Blacklist feature of MercuryS and either query the ip-to-country list at http://countries.nerd.dk/ for uk.countries.nerd.dk and let Mercury add a header on which you can filter later, or, if you want to block foreign IP's at the SMTP level you could set the strictness level in the Blacklist definition setup to 'range', query zz.countries.nerd.dk instead and block everything above and below 127.0.3.58, which is UK.

To mitigate the effect of blocking foreign IP's somewhat you could also set up a Whitelist in MercuryS and query a public whitelist like http://www.dnswl.org/ before.

Best regards

Nico