Qmail Queue Filling up with NDR's

It sounds like your QMail support programs are having trouble with queue management -- specifically, they are not communicating correctly about entries made into the queue.

A little background:
 QMail is not some monolithic program (like sendmail) that does it all in one process thread. Rather, QMail is a collection of programs that communicate with each other to perform specific tasks. (It uses the original UNIX philosophy of making tools: make it do one thing, and do it well).
 One of the most important ways that QMail programs talk to each other is through the queue. In a nutshell though, the message headers and bodies are kept in different files. This way, when I send a message to 10,000 recipients I don't have to write the message body 10,000 times into the queue.

The most likely cause of the behavior you describe is a permissions issue -- either the queue folders don't have the correct permissions, or (more likely) one or more of the QMail programs isn't running with the right permissions (user/group ID).

To debug, we'll need to know the installation method used (who's set of instructions did you follow?). Also, a listing of the running processes (output of [options here are BSD options) ps -efwl). What we're looking for are the permissions on the queue itself, and the user IDs of the processes running QMail.

Or, you could use something like the QMail Toaster's QTP (QMail Toaster Plus) -- which has RPM installations for QMail and a variety of other packages that have a wide history of working well together. While I don't suspect that QMail Toaster (or QTP for that matter) has a BSD install script, you should be able to follow along and make it work on your BSD installation.


Just my thoughts...

Dan
IT4SOHO

Thanks Dan,
Now the queue is just growing.  Nothing is going anywhere, though the log shows it trying.  I don't get that at all...

it didn't like the "f" so I ran "ps -ewl


bash-2.05b# ps -ewl
  UID   PID  PPID CPU PRI NI   VSZ   RSS WCHAN   STAT TT       TIME COMMAND
    0 31097 14405   0  10 20   840   764 wait   I     p0    0:00.04 TERM=xterm SHELL=/usr/local/bin/bash SSH_CLIENT=10.0.0.20 2339
    0 25381 31097   0   2 20   164   436 kqread I+    p0    0:00.43 SHELL=/usr/local/bin/bash TERM=xterm SSH_CLIENT=10.0.0.20 2339
    0 10324 14110   0  10 20   824   752 wait   I     p1    0:00.23 TERM=xterm SHELL=/usr/local/bin/bash SSH_CLIENT=10.0.0.20 2378
    0 25404 10324   0  28 20   340   204 -      R+    p1    0:00.00 SHELL=/usr/local/bin/bash TERM=xterm SSH_CLIENT=10.0.0.20 2378
    0 18642     1  11  18 20   416   312 pause  I     C0-   0:00.01 PATH=/sbin:/bin:/usr/sbin:/usr/bin HOME=/ PWD=/ /bin/sh /comman
    0 30884 18642   0  10 20    76   400 nanosl I     C0-   0:00.18 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 12861 18642   5  -6 20    84   328 piperd I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0  3097 30884   0   2 20    64   392 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0  2681 30884   0   2 20    80   396 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 13701 30884   0   2 20    48   396 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 14657 30884   0   2 20    96   392 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 14323 30884   0   2 20    48   392 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0  9627 30884   0   2 20    76   392 poll   I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 20364 14323   0   2 20   120   424 netcon I     C0-   0:00.64 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 27926 13701  18  18 20   408   308 pause  I     C0-   0:00.01 PATH=/command:/usr/local/bin:/usr/local/sbin:/bin:/sbin:/usr/bi
    0 15821  1917   0   2 20    68   400 select I     C0-   0:00.37 PATH=/var/qmail/bin:/command:/usr/local/bin:/usr/local/sbin:/bi
    0  7344     1   6   3 20    96   476 ttyin  Is+   C0    0:00.01 /usr/libexec/getty Pc ttyC0
    0 26578     1   6   3 20    64   476 ttyin  Is+   C1    0:00.01 /usr/libexec/getty Pc ttyC1
    0  1673     1   6   3 20    84   480 ttyin  Is+   C2    0:00.01 /usr/libexec/getty Pc ttyC2
    0 20222     1   6   3 20    84   476 ttyin  Is+   C3    0:00.01 /usr/libexec/getty Pc ttyC3
    0 32074     1   6   3 20   108   472 ttyin  Is+   C5    0:00.01 /usr/libexec/getty Pc ttyC5
bash-2.05b#

It would help if you could post the exact NDR message that you are getting back. Also make sure your mail server is not an open relay by going to http://www.checkor.com/ and putting in your mail server address. If any tests show an error, you might have spammers using your mail server to send out mail. Are you running any other services on this server, such as any websites? It's possible for hackers to inject mail into your queue using poorly written scripts on your server.

This sounds to me like you are not doing recipient validation. A Spammer will send an email to your server to an alias it knows to NOT EXIST. It will use a faked 'Sender' address and this is the intended recipient. Your server will send an NDR to the faked 'Sender' saying you do not have a user by that name.

The clues here are NDR's to unknown recipients and the From is blank <>. Typical of NDR SPAM.

I don't know about qmail but this will start you off: http://www.lifewithqmail.org/lwq.html#smtp-reject

Shaun

The server is not an open relay.  The messages are about buying watches and enlarging body parts...

I'll check the recipient validation info.  Thanks.