[vchkpw] vpopmail via NFS

2005-06-28 Thread James McMillan

Has anyone tried to hold the vpopmail home as a NFS mount?
I've gotten the mount setup properly, and the permissions are fine, 
however when adding a domain or a user I recieve the following error.


minoru# ./bin/vadduser [EMAIL PROTECTED]
Please enter password for [EMAIL PROTECTED]:
enter password again:
Failed while attempting to add user to auth backend
Error: no auth connection

If copy out all the files in the ~vpopmail mount, unmount, and copy the 
files into the now-local ~/vpopmail I do not have this problem.


Has anyone done this, or know what the problem is?

Thanks in advance.

Jimmy McMillan

--

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106 



[vchkpw] Inter7 announces eMail Messaging Policy Framework (eMPF) open-beta

2005-06-28 Thread Matt Brookings

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Hello, everybody!

eMPF allows mail administrators to set messaging policies for their
local mail users by stating who can message who, and who cant message
who, and by setting domain-wide policies along with exceptions to that
policy.  This patch will of course be GPL, however, because of the
nature of the patch, we are not releasing it for immediate download.

If you are interested in signing up for the open-beta, please be aware
of the following requirements:

Technically savvy enough to apply the patch, and compile with
some very basic instructions

A willingness to provide some feedback on problems encountered,
pros  cons, and implementation ideas

At least one mail server which sends and receives mail for both
internal and external (accounts not hosted locally on the
server) users

As an open-beta user, you should also be aware of the following thing:

This software is beta, and though we've done our best to test
it internally for any problems, it could break.

Because we arent making the software immediately available for
download just yet, we kindly ask open-beta testers not to
re-distribute the patch to others.  However, feel free to pass
this message on to other administrators who may be interested
in being a part of the beta test.

If you're still interested, please subscribe to the eMPF mailing list
by sending an email to [EMAIL PROTECTED], and then send a short
message to the list with basic information about the mail server(s) you
wish to apply the patch to (number of user accounts, approximate mail
volume, etc).

If you're just interested in the development of eMPF, you may of course
subscribe to the list as well.  You arent required to take place in the
beta.

Thanks!
- --
/*
~Matt Brookings [EMAIL PROTECTED]   GnuPG Key 7D7E5F37
~Software developer Systems technician
~Inter7 Internet Technologies, Inc. (815)776-9465
*/
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFCwYEG/pZz8n1+XzcRAmNuAJ9Mu1TOLIyEcxCOOrB3wDfoKH1sAgCfSw/c
0qqfi1qcr439rXUMEeH0uN0=
=EVq2
-END PGP SIGNATURE-


RE: [vchkpw] vpopmail via NFS

2005-06-28 Thread Andre Fortin
I use NFS for my vpopmail directories, and haven't seen this..

I'm assuming you're using CDB files for auth; that means the DB files are
all on NFS.  If you don't have your root mapped properly, you won't be able
to create files.  Try using -maproot=0 (freebsd, others may be similar) in
your exports on the NFS server, and re-mount the NFS mountpoint; see if that
helps.

Andre

 -Original Message-
 From: James McMillan [mailto:[EMAIL PROTECTED]
 Sent: Tuesday, June 28, 2005 11:47 AM
 To: vchkpw@inter7.com
 Subject: [vchkpw] vpopmail via NFS


 Has anyone tried to hold the vpopmail home as a NFS mount?
 I've gotten the mount setup properly, and the permissions are fine,
 however when adding a domain or a user I recieve the following error.

 minoru# ./bin/vadduser [EMAIL PROTECTED]
 Please enter password for [EMAIL PROTECTED]:
 enter password again:
 Failed while attempting to add user to auth backend
 Error: no auth connection

 If copy out all the files in the ~vpopmail mount, unmount, and copy the
 files into the now-local ~/vpopmail I do not have this problem.

 Has anyone done this, or know what the problem is?

 Thanks in advance.

 Jimmy McMillan

 --

 James McMillan
 V.P. Of Information Technology
 www.TheNetMark.com
 412 New Broadway
 Brooklawn, NJ 08030
 888.767.8750 X106





Re: [vchkpw] vpopmail via NFS

2005-06-28 Thread James McMillan

Hey thanks... Let me explain the scenario a bit better.

NFS Server is a Slackware 10.1 box.
NFS Client (and qmail/vpopmail server) is a FreeBSD 5.4 box.

My server side client export file looks like this.

/mnt/hd/mail_store/test minoru(no_root_squash,rw,sync)

I believe the 'no_root_squash' flag is the equiv of the -maproot=0 for 
BSD/other OS's.


I'm not having any issue at all creating files, chowning, or chmoding.  
It seems that it's a simple locking problem.  [lockd] seems to be 
running on the NFS host, however I'm wondering if it's the difference in 
NFS standards.  I've compiled vpopmail with the --disable-file-locking 
configure flag, and then everything seems to work fine, however it's not 
very multi-user-safe.


Any other thoughts would be greatly appreciated.

Jimmy



Andre Fortin wrote:


I use NFS for my vpopmail directories, and haven't seen this..

I'm assuming you're using CDB files for auth; that means the DB files are
all on NFS.  If you don't have your root mapped properly, you won't be able
to create files.  Try using -maproot=0 (freebsd, others may be similar) in
your exports on the NFS server, and re-mount the NFS mountpoint; see if that
helps.

Andre

 


-Original Message-
From: James McMillan [mailto:[EMAIL PROTECTED]
Sent: Tuesday, June 28, 2005 11:47 AM
To: vchkpw@inter7.com
Subject: [vchkpw] vpopmail via NFS


Has anyone tried to hold the vpopmail home as a NFS mount?
I've gotten the mount setup properly, and the permissions are fine,
however when adding a domain or a user I recieve the following error.

minoru# ./bin/vadduser [EMAIL PROTECTED]
Please enter password for [EMAIL PROTECTED]:
enter password again:
Failed while attempting to add user to auth backend
Error: no auth connection

If copy out all the files in the ~vpopmail mount, unmount, and copy the
files into the now-local ~/vpopmail I do not have this problem.

Has anyone done this, or know what the problem is?

Thanks in advance.

Jimmy McMillan

--

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106


   





 




--

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106 



[vchkpw] vpopmail telnet problem

2005-06-28 Thread avudz
Hello,

i have a weird vpopmail instalation, i dont know if this case have
been discuss before, mail can be deliver to mailbox but user can't
pop the message. here is the example  :

[EMAIL PROTECTED] control]# telnet mail.bosung.co.id 110
Trying 202.152.225.122...
Connected to mail.bosung.co.id (202.152.225.122).
Escape character is '^]'.
+OK [EMAIL PROTECTED]
user [EMAIL PROTECTED]
+OK
pass avudz  -- this will stuck after typing pass, no reply with OK

stat -- it work by typing stat after few minutes??
+OK 1 1892

sometimes, it response by File was busy.

here is my pop3 run config :

Connection closed by foreign host.
[EMAIL PROTECTED] control]# cat /var/qmail/supervise/qmail-pop3d/run
#!/bin/sh
PATH=$PATH:/usr/local/bin:/var/qmail/bin
export PATH
MAXPOP3D=30
exec /usr/local/bin/softlimit -m 500 \
tcpserver -H -v -c $MAXPOP3D 0 110 qmail-popup mail.bosung.co.id 
/home/vpopmail/bin/vchkpw \
qmail-pop3d Maildir/ 21

i already use two version of vpopmail :
[EMAIL PROTECTED] src]# ls vpopmail-5.*gz
vpopmail-5.3.27.tar.gz  vpopmail-5.4.2.tar.gz

but it have same problem, please help.

  

-- 
Best regards,
 ./avd  mailto:[EMAIL PROTECTED]



Re: [vchkpw] authdaemond memory leak?

2005-06-28 Thread Billy Newsom

Jan-Willem Regeer wrote:
 Look and see if you have the time to check with valgrind if you can
 find the error. It is in the ports tree, and looks for memory leakage by
 programs.

 Hope you find what the problem is.


 Note: I am not using authdaemond myself.
 Jan-Willem Regeer
 

I tried running it through valgrind's memcheck.  I don't see any issues 
whatsoever to be concerned with.  I ran with it for about 12 hours of 
normal use.  It looks to me that the program itself is collecting a lot 
of information, putting it in memory legitimately, and it simply uses 
gobs of it.  Not a memory leak per se, but a programming mistake.


Just so you know the output I got here was essentially the same when I 
only ran it for a few minutes... the small leaks detected here seem to 
be the same as when I ran the quick tests.


Here's the output I got.  I ran it with the --trace-children=yes option, 
so the process ID's (about three of them) represent the different children.



==45544== Is the main (parent)
==45549== is the worker thread (#2)
==45548== is the worker thread (#1)

Here's the output.
Billy

==45544== Memcheck, a memory error detector for x86-linux.
==45544== Copyright (C) 2002-2004, and GNU GPL'd, by Julian Seward.
==45544== Using valgrind-2.1.2.CVS, a program supervision framework for 
x86-linux.

==45544== Copyright (C) 2000-2004, and GNU GPL'd, by Julian Seward.
==45544==
==45544== My PID = 45544, parent PID = 45543.  Prog and args are:
==45544==/usr/local/libexec/courier-authlib/authdaemond
==45544== For more details, rerun with: -v
==45544==
==45549==
==45548==
==45549== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
==45548== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
==45548== malloc/free: in use at exit: 21793446 bytes in 7800 blocks.
==45548== malloc/free: 39152 allocs, 31352 frees, 150866381 bytes allocated.
==45548== For counts of detected errors, rerun with: -v
==45548== searching for pointers to 7800 not-freed blocks.
==45549== malloc/free: in use at exit: 22053598 bytes in 7893 blocks.
==45549== malloc/free: 39651 allocs, 31758 frees, 152789639 bytes allocated.
==45549== For counts of detected errors, rerun with: -v
==45549== searching for pointers to 7893 not-freed blocks.
==45549== checked 10292868 bytes.
==45548== checked 10195776 bytes.
==45549==
==45549== 11 bytes in 1 blocks are definitely lost in loss record 1 of 12
==45548==
==45548== 11 bytes in 1 blocks are definitely lost in loss record 1 of 12
==45549==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)
==45548==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)

==45549==by 0x3C1038A2: strdup (in /lib/libc.so.5)
==45548==by 0x3C1038A2: strdup (in /lib/libc.so.5)
==45549==by 0x8049963: (within 
/usr/local/libexec/courier-authlib/authdaemond)
==45548==by 0x8049963: (within 
/usr/local/libexec/courier-authlib/authdaemond)
==45549==by 0x804B014: start (in 
/usr/local/libexec/courier-authlib/authdaemond)
==45548==by 0x804B014: start (in 
/usr/local/libexec/courier-authlib/authdaemond)

==45548==
==45549==
==45548==
==45549==
==45548== 34 bytes in 2 blocks are definitely lost in loss record 4 of 12
==45549== 34 bytes in 2 blocks are definitely lost in loss record 4 of 12
==45548==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)
==45549==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)

==45548==by 0x3C03CA4C: lt_emalloc (in /usr/local/lib/libltdl.so.4)
==45549==by 0x3C03CA4C: lt_emalloc (in /usr/local/lib/libltdl.so.4)
==45548==by 0x3C03D6F2: canonicalize_path (in 
/usr/local/lib/libltdl.so.4)
==45549==by 0x3C03D6F2: canonicalize_path (in 
/usr/local/lib/libltdl.so.4)

==45548==by 0x3C03E494: try_dlopen (in /usr/local/lib/libltdl.so.4)
==45549==by 0x3C03E494: try_dlopen (in /usr/local/lib/libltdl.so.4)
==45548==
==45548==
==45549==
==45548== 455840 bytes in 2590 blocks are possibly lost in loss record 
11 of 12

==45549==
==45549== 461296 bytes in 2621 blocks are possibly lost in loss record 
11 of 12
==45548==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)
==45549==at 0x3C03772F: malloc (in 
/usr/local/lib/valgrind/vgpreload_memcheck.so)
==45548==by 0x3C2888CE: my_malloc (in 
/usr/local/lib/mysql/libmysqlclient.so.14)
==45549==by 0x3C2888CE: my_malloc (in 
/usr/local/lib/mysql/libmysqlclient.so.14)
==45548==by 0x3C2A3536: mysql_store_result (in 
/usr/local/lib/mysql/libmysqlclient.so.14)
==45549==by 0x3C2A3536: mysql_store_result (in 
/usr/local/lib/mysql/libmysqlclient.so.14)
==45548==by 0x3C26CE5A: vget_limits (in 
/usr/local/lib/courier-authlib/libauthvchkpw.so)
==45549==by 0x3C26CE5A: vget_limits (in 
/usr/local/lib/courier-authlib/libauthvchkpw.so)

==45548==
==45548== LEAK SUMMARY:
==45548==definitely lost: 45 bytes in 3 blocks.
==45549==
==45548==possibly 

[vchkpw] vpopmail ldap round robin

2005-06-28 Thread Mathieu CHATEAU
Hello all,

some points regarding vpopmail  ldap.

1/ qmailUser.schema

   while it is working with openldap 2.2.26, it's not with openldap
   2.2.27.
  it inherits objects from person objectclass where they are
  MUST. qmailUser call them in a MAY but it's not correct.
  We can go from may to must but not the oppposite.

  The matter is that qmailadmin does not create the sn  cn
  and so it's not possible to create account with 2.2.27

2/ Round robin DNS
 i have put ldap.mydomain.com in vldap.h it returns two ip in
 round robin. it works, but if i shutdown the one used, vchkpw
 does not fail over to the other one.
 it should go to the other one after the dns ttl expires. does
 someone already done it with success ?
 how to clear the dns cache entry used by vpopmail ?


thank you in advance,

  

-- 
Best regards,
 Mathieu  mailto:[EMAIL PROTECTED]



RE: [vchkpw] vpopmail via NFS

2005-06-28 Thread Clayton Weise
I would reccomend having your slackware server running MySQL or some
other type of database server.  The installing the mysql client binaries
and libraries on your frontend mail servers.  Compile vpopmail for
mysql (or the db backend of your choice) with replication.  Designate
your slackware server as your writing server, and the local boxes as
your reading server.

Then, on each local machine setup mysql replication from the frontend
mail server(s) as slaves to the master slackware server on the backend.
Then, mount your ~vpopmail/etc and ~vpopmail/domains, and possibly
~vpopmail/bin depending on how you want to set it up, via NFS from the
slackware backend server to your bsd frontend servers.

With a SQL server handling the authentication it will help you in a
couple of ways.  First is that you won't need to worry about file
locking issues since the the SQL server will take care of all of that.
Second would be improved performance, since you'll be querying a local
database instead of having to travel across NFS to deliver the mail.
And lastly, you could even apply the MySQL patch to tcpserver to do
pop-before-smtp authentication out of the MySQL database in place of the
cdb file.  This would only be a problem if you were getting so many pop
connections that it couldn't rebuild the cdb file fast enough.  Another
workaround for the whole rebuilding tcp.smtp.cdb file thing is to use
the relay-ctrl program from untroubled.org for relay control.

-Clayton

-Original Message-
From: James McMillan [mailto:[EMAIL PROTECTED] 
Sent: Tuesday, June 28, 2005 10:39 AM
To: vchkpw@inter7.com
Subject: Re: [vchkpw] vpopmail via NFS

Hey thanks... Let me explain the scenario a bit better.

NFS Server is a Slackware 10.1 box.
NFS Client (and qmail/vpopmail server) is a FreeBSD 5.4 box.

My server side client export file looks like this.

/mnt/hd/mail_store/test minoru(no_root_squash,rw,sync)

I believe the 'no_root_squash' flag is the equiv of the -maproot=0 for 
BSD/other OS's.

I'm not having any issue at all creating files, chowning, or chmoding.  
It seems that it's a simple locking problem.  [lockd] seems to be 
running on the NFS host, however I'm wondering if it's the difference in

NFS standards.  I've compiled vpopmail with the --disable-file-locking 
configure flag, and then everything seems to work fine, however it's not

very multi-user-safe.

Any other thoughts would be greatly appreciated.

Jimmy



Andre Fortin wrote:

I use NFS for my vpopmail directories, and haven't seen this..

I'm assuming you're using CDB files for auth; that means the DB files
are
all on NFS.  If you don't have your root mapped properly, you won't be
able
to create files.  Try using -maproot=0 (freebsd, others may be similar)
in
your exports on the NFS server, and re-mount the NFS mountpoint; see if
that
helps.

Andre

  

-Original Message-
From: James McMillan [mailto:[EMAIL PROTECTED]
Sent: Tuesday, June 28, 2005 11:47 AM
To: vchkpw@inter7.com
Subject: [vchkpw] vpopmail via NFS


Has anyone tried to hold the vpopmail home as a NFS mount?
I've gotten the mount setup properly, and the permissions are fine,
however when adding a domain or a user I recieve the following error.

minoru# ./bin/vadduser [EMAIL PROTECTED]
Please enter password for [EMAIL PROTECTED]:
enter password again:
Failed while attempting to add user to auth backend
Error: no auth connection

If copy out all the files in the ~vpopmail mount, unmount, and copy
the
files into the now-local ~/vpopmail I do not have this problem.

Has anyone done this, or know what the problem is?

Thanks in advance.

Jimmy McMillan

--

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106







  



-- 

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106 



RE: [vchkpw] vpopmail via NFS

2005-06-28 Thread Clayton Weise
Slight correction on my verbage there:

Second would be improved performance, since you'll be querying a local
database instead of having to travel across NFS to deliver the mail.

I meant to say that it won't have to travel across NFS to find if the
user exists, only to deliver the mail.



Re: [vchkpw] vpopmail via NFS

2005-06-28 Thread James McMillan
Clayton, thanks... that's exactully what i did.  Everything seems good 
now.  Thanks a million for you input.


Jimmy

Clayton Weise wrote:


Slight correction on my verbage there:

 


Second would be improved performance, since you'll be querying a local
database instead of having to travel across NFS to deliver the mail.
   



I meant to say that it won't have to travel across NFS to find if the
user exists, only to deliver the mail.



 




--

James McMillan
V.P. Of Information Technology
www.TheNetMark.com
412 New Broadway
Brooklawn, NJ 08030
888.767.8750 X106