[vchkpw] Permissions on vpopmail install suggestion

2007-08-22 Thread Bob
Many of us use either qmail-scanner-queue or simscan via patched qmail 
ahead of vpopmail. If, as is good prcatice, we allow the scanner to run 
under its own user ID, vchkpw will fail because instead of running as 
user vchkpw it is running as the scanner user which doesn't have 
access to the password files. I would like to suggest that in the make 
install, the permission for vchkpw be set to 4711 so that it will always 
execute as the vchkpw user. Doing this will eliminate a bit of extra 
work when upgrading and will stop the large number of user questions 
when they do their qmail installs.


Thanks,
Bob


Re: [vchkpw] vpopmail 5.4.20 released

2007-08-22 Thread Steve Cole
On Wednesday 22 August 2007, Rick Widmer wrote:
 http://vpopmail.sf.net/

 5.4.20 - released 21-Aug-07

Feels odd to ask this, but in the UPGRADE document, it lists the pw_domain 
columns as requiring updates, but my databases have just domain in them 
except for the vpopmail table.

I've upgraded them manually, but should I be re-naming the columns manually?   
I'd prefer not to take the plunge and ask instead. :)

Moving from 5.4.17.

Thanks for any help.

-- 
Cheers,
Steve  |President  Systems Administrator,  Kingston Online Services
   |URL: http://www.kos.net/   613-549-8667/613-968-7137


[vchkpw] domains spanning directories

2007-08-22 Thread Trey Nolen
When new users are added to vpopmail and you reach a certain number of them,
it creates a new folder and puts users in that folder. Then, when you reach
another level of users, it adds a new folder and repeats the cycle.

Presumably, it does the same thing with domains, but we have 25 domains at
the top level, then a 0 directory with 190 domains. It seem that it should
have spanned to a 1 directory by now.  Is this the case?  If not, how can
I fix it so that it will create a 1 and move on?   I don't really see a
need to relocate the existing ones...


Trey Nolen




Re: [vchkpw] Permissions on vpopmail install suggestion

2007-08-22 Thread John Simpson

On 2007-08-22, at 1534, Bob wrote:


Many of us use either qmail-scanner-queue or simscan via patched  
qmail ahead of vpopmail. If, as is good prcatice, we allow the  
scanner to run under its own user ID, vchkpw will fail because  
instead of running as user vchkpw it is running as the scanner  
user which doesn't have access to the password files. I would like  
to suggest that in the make install, the permission for vchkpw be  
set to 4711 so that it will always execute as the vchkpw user.  
Doing this will eliminate a bit of extra work when upgrading and  
will stop the large number of user questions when they do their  
qmail installs.


i've been doing this for several years.

and for those who run into this issue, at least with simscan, i refer  
you to this page which explains the problem and how to work around it:


http://qmail.jms1.net/simscan/troubleshooting.shtml

i also think the install procedure for simscan needs to be changed a  
bit, but that probably belongs on the simscan mailing list.



| John M. Simpson---   KG4ZOW   ---Programmer At Large |
| http://www.jms1.net/ [EMAIL PROTECTED] |

| http://video.google.com/videoplay?docid=-1656880303867390173 |





PGP.sig
Description: This is a digitally signed message part


Re: [vchkpw] Permissions on vpopmail install suggestion

2007-08-22 Thread John Simpson

On 2007-08-22, at 1952, John Simpson wrote:

On 2007-08-22, at 1534, Bob wrote:


Many of us use either qmail-scanner-queue or simscan via patched  
qmail ahead of vpopmail. If, as is good prcatice, we allow the  
scanner to run under its own user ID, vchkpw will fail because  
instead of running as user vchkpw it is running as the scanner  
user which doesn't have access to the password files. I would like  
to suggest that in the make install, the permission for vchkpw be  
set to 4711 so that it will always execute as the vchkpw user.  
Doing this will eliminate a bit of extra work when upgrading and  
will stop the large number of user questions when they do their  
qmail installs.


i've been doing this for several years.


let me correct this statement... i DID this for years, but i don't do  
it any longer.


the problem that bob is talking about is this- if somebody is using  
the normal AUTH patch for qmail, and wants to use vchkpw as a  
method of allowing qmail-smtpd to validate AUTH commands, the  
vchkpw command needs to have permission to read the vpasswd.cdb  
files. and if it runs as the qmaild user, it doesn't have that  
permission.


one solution, and what i did myself for a few years, is to make  
vchkpw run setuid root. however, some people set up vpopmail  
domains using different system uid's (i.e. vadddomain with the -u  
option) for different domains, as a way to implement domain quotas  
by setting a filesystem quota on the uid which controls the domain.  
in this situation, you do NOT want vchkpw to be setuid to the  
vpopmail user.


a better solution is to make qmail-smtpd use something other than a  
checkpassword program (which is what vchkpw actually is) to verify  
passwords. this was the reason that i wrote an addition to my  
combined qmail patch, which teaches qmail-smtpd to use an auth.cdb  
file, with email addresses as keys and encrypted passwords as values,  
to validate AUTH commands. since i've started using this, i haven't  
needed vchkpw to be setuid, and in fact it's not setuid on my  
server any more.


one of these days i'll get around to writing an AUTH_CDB patch for  
djb's virgin qmail-1.03 code, and probably for netqmail-1.05 as well,  
but for now it's available in my combined patch (which has lots of  
other yummy features as well.)



| John M. Simpson---   KG4ZOW   ---Programmer At Large |
| http://www.jms1.net/ [EMAIL PROTECTED] |

| http://video.google.com/videoplay?docid=-1656880303867390173 |





PGP.sig
Description: This is a digitally signed message part